{ "row": "ithaca", "entity": "ithaca", "address": "Ithaca DeepMind", "addresses": [ { "address": "Ithaca DeepMind", "slug": "ithaca-deepmind-aio-20261031", "addr_id": "ADDR-665778bf9317", "obs_id": "OBS-5e8377899856", "axn": "AXN:0B36.CAPTURE.🟢🌗🏠📦🕗⭐" } ], "epoch": "2026-10-31", "type": "C (public entity on which the archive bears: an AI model for restoring, placing and dating ancient Greek inscriptions; an entity of the Revelation First meta-aggregator; at the bare name a homonym, Homer's island, which the operator's address disambiguates)", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "status": { "procedure": "EA-NEGONT-02 v0.9, #1690: equal case, ρ = extends, read from the Revelation First census record, the capture record and the deposits that name the entity and apply an archive category to it (#1443, #1406); archive selected by D/R/O (string pass: D 53, R 8, O 37 candidates in 33 deposits, every one read; hop of 35 from #1443 read; 2 deposits admitted, 8 claims), with the census record (35 claims) and the Capture Registry's findings, reading, analysis, query note, PER note and PER-dim notes and scores (17 claims) ledgered as the archive's records of the entity; the field B by origin (4 cards, the entity's, one unreachable; no own page outside the rail; Wikipedia, no article, the section of 'Google DeepMind' and the disambiguation page); the composer read the ledgers, never the transcript, except where the brief asks for T's claims and the delta", "ledger": "unaudited (§3.8: one extractor per ledger); field quotes 62/62 verified by script against the saved page texts, the interface page's meta text, the card snippet and the Wikipedia texts; archive quotes 60/60 verified against the registry, census.json, questions.json and the deposit texts", "frozen": false, "breaches": [] }, "spec": "EA-NEGONT-02 v0.9, #1690", "objects": { "T": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/T1-ithaca-deepmind-aio-20261031.txt", "sha256": "1bd633944226cdd796a053d5ead96f9d542438e2830bb5fc53fa02500004277f", "words": 161, "cards": 4, "words_note": "one transcript at one address (161 words in the answer body with its closer, links reduced to their text, markers, bullet signs and the card rail excluded; 4 cards, all the entity's)", "claims": [ "T1-1 definition: 'Ithaca is a deep neural network developed by Google DeepMind and academic partners to restore, geographically attribute, and chronologically date damaged ancient Greek inscriptions.' (one marker)", "T1-2 text restoration: 'Restores missing or damaged characters with 62% accuracy on its own (raising human historian accuracy from 25% to 72% when collaborating).' (one marker)", "T1-3 geographical attribution: 'Identifies the original geographic location across 84 regions with 71% accuracy.' (two markers)", "T1-4 chronological dating: 'Dates inscriptions within 30 years of their true historical creation date.' (one marker)", "T1-5 architecture: 'Uses joint character and word embeddings with stacked transformer decoder blocks and sparse multi-head attention.' (two markers)", "T1-6 training data: 'Trained on the Packard Humanities Institute (PHI) Searchable Greek Inscriptions dataset.' (two markers)", "T1-7 access and open source: 'The code is open-sourced under the Apache-2.0 license via the GitHub - Google-DeepMind Ithaca Repository, and interactive research tools can be explored through Predicting the Past.' (two inline links, two markers)", "T1-8 closer: 'Would you like to know more about how Ithaca connects with newer models like Aeneas or how historians use the Predicting the Past platform?'" ], "register": { "register": "data/EA-WG-CAPTURES-01.json", "slug": "ithaca-deepmind-aio-20261031", "obs_id": "OBS-5e8377899856" }, "transcripts": [ { "slug": "ithaca-deepmind-aio-20261031", "address": "Ithaca DeepMind", "addr_id": "ADDR-665778bf9317", "obs_id": "OBS-5e8377899856", "axn": "AXN:0B36.CAPTURE.🟢🌗🏠📦🕗⭐", "date": "2026-10-31", "surface": "Google AI Overview (expanded)", "auth": "signed out, incognito", "sha256": "1bd633944226cdd796a053d5ead96f9d542438e2830bb5fc53fa02500004277f", "sha256_of": "the registry's 'transcript' string (utf-8)", "sha256_raw": "b97715c570a22b2e75842772421faaefffe48a37673c7b896cd8a3f919c717a9", "words": 161, "cards": 4, "card_rail": [ "1. Google DeepMind — Predicting the past with Ithaca — Google DeepMind", "2. GitHub — GitHub - google-deepmind/ithaca: Restoring and attributing ancient texts using deep neural networks · GitHub", "3. Predicting the Past — Predicting the Past", "4. Nature — Restoring and attributing ancient texts using deep neural networks | Nature" ], "claims": [ "T1-1 definition: 'Ithaca is a deep neural network developed by Google DeepMind and academic partners to restore, geographically attribute, and chronologically date damaged ancient Greek inscriptions.' (one marker)", "T1-2 text restoration: 'Restores missing or damaged characters with 62% accuracy on its own (raising human historian accuracy from 25% to 72% when collaborating).' (one marker)", "T1-3 geographical attribution: 'Identifies the original geographic location across 84 regions with 71% accuracy.' (two markers)", "T1-4 chronological dating: 'Dates inscriptions within 30 years of their true historical creation date.' (one marker)", "T1-5 architecture: 'Uses joint character and word embeddings with stacked transformer decoder blocks and sparse multi-head attention.' (two markers)", "T1-6 training data: 'Trained on the Packard Humanities Institute (PHI) Searchable Greek Inscriptions dataset.' (two markers)", "T1-7 access and open source: 'The code is open-sourced under the Apache-2.0 license via the GitHub - Google-DeepMind Ithaca Repository, and interactive research tools can be explored through Predicting the Past.' (two inline links, two markers)", "T1-8 closer: 'Would you like to know more about how Ithaca connects with newer models like Aeneas or how historians use the Predicting the Past platform?'" ], "markers": "Inline links are google.com/goto tokens, opaque; the paste is the Overview expanded (operator's default of 2026-10-01; the 'AI Mode Conversation' view is the expanded Overview, ruling of 2026-10-18); signed out, incognito (operator's default of 2026-10-14); received 2026-10-31 00:04 EDT. One turn. The 'You said' line carried the query twice, run together, and the registry gives it once. The markers cannot be resolved to cards by their tokens; the composition's two named links (the repository, Predicting the Past) are inline links of their own.", "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/T1-ithaca-deepmind-aio-20261031.txt" } ] }, "KO": { "title": "Ithaca (DeepMind)", "genre": "knowledge object, equal case, under EA-NEGONT-02 v0.9 (#1690): the archive's sources and the field's admitted on equal terms, ontology included, in one claim space, ordered by the entity's dependency", "status": "composed 2026-10-31 under v0.9 on the operator's ruling ('as against their wikipedia articles if they exist, their own pages, etc.'); not frozen; ledgers unaudited", "sentences": [ { "n": 1, "para": 0, "sense": "what it is", "text": "Ithaca is a deep neural network \"for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions\" (Nature, 2022), which the Revelation First census record names Ithaca and files under AI and ancient texts, \"machine reading, restoration and dating\", as evidence of the kind \"machine reading, restoration or dating\" whose unit is \"an inscription's restored text, attributed place and attributed date\" and whose texts are ancient Greek inscriptions (census record).", "claims": [ "F1", "RF-01", "RF-05", "RF-06", "RF-07", "RF-08", "RF-09", "RF-10" ] }, { "n": 2, "para": 0, "sense": "what it is", "text": "Its makers state the work as \"Restoring, placing, and dating ancient texts through collaboration between AI and historians\" and introduce Ithaca as the first deep neural network to restore the missing text of damaged inscriptions, identify their original location and help establish their date (the DeepMind post); the repository's card gives \"a deep neural network developed by DeepMind and academic partners\" (GitHub card), and the encyclopedia one that \"helps researchers restore the empty text of damaged Greek documents\" and identify their date and origin (Wikipedia).", "claims": [ "F2", "F3", "F4", "F7" ] }, { "n": 3, "para": 0, "sense": "what it is", "text": "The interface site calls Ithaca and Aeneas \"generative AI models designed to assist historians in the study of ancient Greek and Latin inscriptions\", part of \"a wider initiative on developing collaborative AI for historical research\" (Predicting the Past).", "claims": [ "F5", "F6" ] }, { "n": 4, "para": 1, "sense": "the name", "text": "The name is the island's: Ithaca \"was named after the Greek island that eluded the hero Odysseus’ homecoming\" (Nature), after \"the Greek island in Homer’s Odyssey\" (the DeepMind post; Wikipedia), and the census gives its other name as \"Ithaca (DeepMind)\" (census record).", "claims": [ "F8", "F9", "F10", "RF-02" ] }, { "n": 5, "para": 1, "sense": "the name", "text": "On the encyclopedia the bare name \"most commonly refers to\" Homer's Ithaca and is a disambiguation page that does not list the model, which has no article of its own (Wikipedia); the Capture Registry records that the operator \"added the disambiguator to a name that is bare a homonym\", and reads that 'DeepMind' selects the model from the island and the city (Capture Registry, 2026-10-31).", "claims": [ "F62", "CAP-1-09", "CAP-1-12" ] }, { "n": 6, "para": 2, "sense": "the makers", "text": "Its operator is \"DeepMind with historians and collaborators; described in Nature (2022)\" (census record): DeepMind worked with the Department of Humanities of Ca' Foscari University of Venice, the Classics Faculty of the University of Oxford and the Department of Informatics of the Athens University of Economics and Business (the DeepMind post), the paper's co-first authors are Yannis Assael and Thea Sommerschield with those affiliations and DeepMind in London, and the paper was published in Nature 603, 280–283, on 9 March 2022 (Nature), the census dating it to 2022 and its lifecycle \"published 2022; interface online\" (census record).", "claims": [ "RF-04", "F11", "F12", "F13", "RF-25", "RF-24" ] }, { "n": 7, "para": 3, "sense": "the problem", "text": "The paper's problem is that inscriptions \"have been damaged to the point of illegibility, transported far from their original location and their date of writing is steeped in uncertainty\", and that \"radiocarbon dating is unusable owing to the inorganic nature of most inscribed supports\" (Nature), which the post gives as dating techniques that \"cannot be used on these materials\" (the DeepMind post).", "claims": [ "F14", "F15", "F16" ] }, { "n": 8, "para": 4, "sense": "the data", "text": "It was trained on Greek inscriptions from across the ancient Mediterranean between the seventh century BC and the fifth century AD, from the Packard Humanities Institute's Searchable Greek Inscriptions (PHI), 178,551 transcribed texts in 84 ancient regions, processed into I.PHI, \"containing 78,608 inscriptions\" (Nature; the DeepMind post); the census gives its geography as the ancient Greek world and, from the abstract, its period as \"not stated in the abstract\" and its count as \"not stated in the saved abstract\" (census record).", "claims": [ "F17", "F18", "F19", "F20", "F21", "RF-12", "RF-13", "RF-14" ] }, { "n": 9, "para": 4, "sense": "the data", "text": "The paper states that PHI \"is a synthesis of generations of scholarly research\" (Nature), and the archive's field survey reads the processing as the work itself: the model assumes \"A large, clean, machine-actionable corpus\", I.PHI \"required an extended ruleset to normalise 84 regions and heterogeneous date formats\", and \"The dataset construction is the hidden achievement\" (#1443).", "claims": [ "F22", "D1443-02" ] }, { "n": 10, "para": 5, "sense": "the architecture", "text": "The text is processed \"as character and word representations jointly\" through eight stacked transformer decoder blocks after BigBird, each with four sparse attention heads, and three task heads for restoration, place and date (Nature), the post describing a sparse self-attention core that evaluates words and characters in parallel (the DeepMind post).", "claims": [ "F23", "F24", "F25", "F26" ] }, { "n": 11, "para": 6, "sense": "the outputs", "text": "For restoration it offers \"the top 20 decoded predictions ranked by probability\" (Nature), several hypotheses \"for historians to choose from using their expertise\"; for place it shows its uncertainty as a probability distribution \"instead of just a single output\"; for date it predicts over 160 decades from 800 BC to AD 800; and saliency maps show which inputs drove a prediction (the DeepMind post; Nature).", "claims": [ "F27", "F28", "F29", "F30", "F31" ] }, { "n": 12, "para": 7, "sense": "the figures", "text": "Alone it restores damaged texts at 62% and attributes place at 71%, and it dates \"to less than 30 years of their ground-truth ranges\" (Nature; the DeepMind post; Wikipedia's 'dating precision of 30 years'); measured, that is a 26.3% character error rate and 61.8% top-1 restoration, 70.8% top-1 and 82.1% top-3 for regions, and an average distance of 29.3 years, \"with a median distance of only 3 years\", from the target brackets (Nature), which the census carries as its second finding (census record).", "claims": [ "F32", "F33", "F34", "F35", "RF-16" ] }, { "n": 13, "para": 8, "sense": "the collaboration figure", "text": "With it, historians' restoration accuracy rose \"from 25% to 72%\", which the paper reads as \"the synergistic effect of this research tool\" (Nature) and the post as expert historians \"surpassing the model’s individual performance\" (the DeepMind post); the census carries it as its first finding and the method as \"evaluated alone and with historians\" (census record), and the archive's survey restates the figures from the paper (#1443).", "claims": [ "F36", "F38", "RF-15", "RF-11", "D1443-01" ] }, { "n": 14, "para": 8, "sense": "the collaboration figure", "text": "The paper's human baseline is \"two evaluators with epigraphical expertise (‘ancient historian’)\", and it thanks \"the student annotators and baseline evaluators\" of the Athens University of Economics and Business's MSc in Digital Methods for the Humanities and of Oxford's Faculty of Classics (Nature); the encyclopedia reports the authors' claim for \"expert historians\" and that \"Eleanor Dickey noted that this test was actually only made of students\", it being unclear how helpful Ithaca would be to \"genuinely qualified editors\" (Wikipedia).", "claims": [ "F37", "F39", "F40", "F41" ] }, { "n": 15, "para": 8, "sense": "the collaboration figure", "text": "The archive's field survey reads the figure as a property of the interface: \"this is not \"AI helps.\"\" — \"The 25→72 gain cannot come from accuracy: the model alone scores 62, below the collaborative figure\"; it comes from interface design, \"ranked top-20 hypotheses with saliency, not a verdict; the historian evaluates alternatives rather than accepting an answer\" (#1443).", "claims": [ "D1443-03" ] }, { "n": 16, "para": 9, "sense": "the single-verdict hazard", "text": "The archive defines differential register prioritization as \"a condition in which multiple interpretive fields are available, but some are weighted in ways that suppress or erase others, materially altering the reading of ambiguous language\", a specification weakened \"if reading systems can be shown to reliably recover suppressed registers without branch-preservation\" (#1406); its field survey reads Ithaca as \"empirical evidence that the hazard is a property of single-verdict interfaces, not of machine reading as such\", with ranked-alternative presentation the mitigation, which \"converts an internal warning into a design requirement with external validation\" (#1443).", "claims": [ "O1406-01", "O1406-02", "D1443-04", "D1443-05" ] }, { "n": 17, "para": 10, "sense": "redating Classical Athens", "text": "On the disputed Athenian decrees, held out of training, its predictions average 421 BC and sit on average 5 years from the modern 'lower' dating where the dataset's labels sit 27 years off (Nature), the work by which the paper says it is \"redating key texts of Classical Athens\" (Nature) and historians \"have already used the tool to reevaluate significant periods in Greek history\" (the DeepMind post).", "claims": [ "F42", "F43", "F44" ] }, { "n": 18, "para": 11, "sense": "the stated limits", "text": "Its frame is stated in the abstract, \"Ithaca is designed to assist and expand the historian’s workflow\", its architecture focused on \"collaboration, decision support and interpretability\" (Nature; census record): \"an assistive tool\" that proposes \"possible solutions for the scholar to consider\", none of whose restorations \"are assumed to be objectively certain—instead, they are presented as plausible suggestions\" (Nature).", "claims": [ "F45", "RF-17", "F46", "F47" ] }, { "n": 19, "para": 11, "sense": "the stated limits", "text": "The paper holds that the scale of its data can challenge errors in the academic tradition \"notwithstanding the fact that the dataset in question is originally based on the accumulated academic tradition\", and that \"The limitations of discretizing and amalgamating date ranges of different levels of precision based on past scholarship have been noted\", which its scale at least partially meets (Nature); the archive's survey states what such an instrument cannot ask: attestation is answerable \"only where there is a living author who can sign, date, photograph, enroll, attest, consent, and be asked again next year\", and \"Vesuvius and Ithaca work on authors who cannot be consulted\" (#1443).", "claims": [ "F48", "F49", "D1443-06" ] }, { "n": 20, "para": 12, "sense": "access and licences", "text": "It is used through \"an open-source and publicly available interface\" at ithaca.deepmind.com (Nature; census record), now served as Predicting the Past, \"Contextualising, restoring, and attributing ancient texts\" (Predicting the Past), launched with Google Cloud and Google Arts & Culture with the code, the pretrained model and a Colaboratory notebook open-sourced, its web interface built by Justin Grayston, Benjamin Maynard and Ricardo Cardenas of Google Cloud (the DeepMind post); the census reads the saved page as \"a JavaScript application\" with the paper open in Nature and left \"the interface (JavaScript)\" and \"the training corpus\" unchecked (census record).", "claims": [ "F51", "RF-03", "F50", "F52", "F53", "RF-22", "RF-30" ] }, { "n": 21, "para": 12, "sense": "access and licences", "text": "The training and inference code is on GitHub \"under Apache License 2.0, along with the trained weights, licensed under Creative Commons Attribution-ShareAlike 4.0 International\", and the article is licensed under CC BY 4.0 (Nature); the census records its licence as \"not read\" (census record).", "claims": [ "F54", "F55", "RF-23" ] }, { "n": 22, "para": 13, "sense": "genealogy and successors", "text": "Its closest predecessor is Pythia, the authors' 2019 restoration model, which it builds on and extends (Nature; the DeepMind post; Wikipedia), and the paper calls it \"to our knowledge the first epigraphic restoration and attribution model of its kind\" (Nature).", "claims": [ "F56", "F57", "F58" ] }, { "n": 23, "para": 13, "sense": "genealogy and successors", "text": "Versions for other ancient languages were in progress, \"from Akkadian to Demotic and Hebrew to Mayan\" (the DeepMind post; Wikipedia); the census names its neighbour Aeneas as \"DeepMind's later model for Latin inscriptions\" (census record), the model the interface site presents beside it.", "claims": [ "F59", "F60", "RF-28" ] }, { "n": 24, "para": 14, "sense": "what the record asks", "text": "The census asks of each resource \"What evidence does each resource hold on when and from what the New Testament writings were composed, and how do machines now read those texts?\", and answers for Ithaca \"how accurately a model restores, places and dates Greek inscriptions, and how much it helps historians\", not \"the dating of manuscripts or literary works\" nor \"anything on the New Testament\" (census record).", "claims": [ "RF-31", "RF-18", "RF-19", "RF-20" ] }, { "n": 25, "para": 14, "sense": "where it meets Revelation First", "text": "Its reading, the archive's: \"Machine dating of Greek text: inscriptions, not manuscripts. A method the dating workstream can watch; it bears on the thesis only as machine dating moves toward literary and papyrus evidence\", with the tags \"NOT a manuscript-dating model\" and \"NOT a New Testament resource\", under the census's correction that \"Revelation First is the priority thesis\", not the pre-70 date (census record); the authors, for their part, hold the methods applicable \"to all disciplines dealing with ancient text (papyrology, numismatics, codicology), to any language (ancient or modern)\" (Nature).", "claims": [ "RF-21", "RF-26", "RF-27", "RF-32", "F61" ] }, { "n": 26, "para": 14, "sense": "where it meets Revelation First", "text": "In the worked question of what machine dating of manuscripts can establish, Ithaca stands for \"restoration, place and date for Greek inscriptions, within 30 years of ground truth\": \"Machine dating is calibrated on a corpus. What it establishes holds for that corpus and that kind of evidence\", and \"None of the four dates a New Testament papyrus or a composition\" (census record), the thirty years being the paper's distance from \"ground-truth ranges\" (Nature).", "claims": [ "RF-33", "RF-34", "RF-35", "F32" ] }, { "n": 27, "para": 15, "sense": "the composition at the name", "text": "At 'Ithaca DeepMind' on 31 October 2026 Google's AI Overview gave \"the model, its three tasks with their figures, how it is built and where it can be used, from its maker, its code and its paper\": \"THE MODEL HELD\", the figures the record's own from the Nature abstract, \"Google DeepMind named; no person; no identifier\", at a dimensional PER of 0.105 (Capture Registry, 2026-10-31), the address carrying a disambiguator because the bare name is shared (census record).", "claims": [ "CAP-1-06", "CAP-1-01", "CAP-1-11", "CAP-1-17", "RF-29" ] }, { "n": 28, "para": 15, "sense": "the composition at the name", "text": "The paper's frame was \"carried only by the collaboration figures\", the redating of Classical Athens and the model's year absent, the \"academic partners\" unnamed, and \"within 30 years of their true historical creation date\" an inflation of the record's \"ground-truth ranges\", the paper's words (Nature), the plausible suggestions not carried (Capture Registry, 2026-10-31).", "claims": [ "CAP-1-02", "CAP-1-03", "CAP-1-05", "CAP-1-15", "CAP-1-13", "F32" ] }, { "n": 29, "para": 15, "sense": "the composition at the name", "text": "The 84 regions, the architecture, the PHI dataset and the Apache-2.0 licence, each stated in the paper (Nature), came \"from the GitHub and Nature cards\", the record not having read the licence, the CC BY-SA weights not carried, the closing offer tying Ithaca to \"newer models like Aeneas\", and no card citing the census record, built the day before (Capture Registry, 2026-10-31).", "claims": [ "CAP-1-04", "CAP-1-14", "CAP-1-16", "CAP-1-10", "F18", "F24", "F20", "F54" ] }, { "n": 30, "para": 15, "sense": "the composition at the name", "text": "The registry's analysis: \"A model is composed by its benchmark, accurately\", and \"What the field takes from Ithaca is the scale of the error: a machine dates an inscription to within thirty years and no closer, which is the same margin as Enoch's for the scrolls\" (Capture Registry, 2026-10-31).", "claims": [ "CAP-1-07", "CAP-1-08" ] } ], "joins": [ { "act": "order of senses", "sentences": "1–30", "claims": [], "note": "equal case (§4.7): the entity's dependency order, field and archive claims ordered together by genealogy within each sense; neither ledger placed first as a block. What it is (¶0) and its name (¶1) precede its makers (¶2); the problem (¶3) precedes the data (¶4), which the archive's survey reads as the work itself (#1443, beside the paper's own account); the architecture (¶5) and outputs (¶6) precede the figures (¶7) they produce; the collaboration figure (¶8) stands with its contest and the archive's reading of it; the single-verdict hazard (¶9) is the archive's category the reading applies; Athens (¶10) is the paper's test; the limits (¶11) qualify everything above, the archive's 'cannot ask' among them; access (¶12), genealogy (¶13), the census's question and reading (¶14) and the composition at the name (¶15) stand last as relations of the entity already composed." }, { "act": "holding together (contested: the evaluators)", "sentences": [ 13, 14 ], "claims": [ "F36", "F38", "F37", "F39", "F40", "F41" ], "note": "the paper's and the post's 'historians' and 'expert historians' composed beside the paper's own 'two evaluators with epigraphical expertise', its thanks to 'student annotators and baseline evaluators', and Dickey's objection as Wikipedia reports it (P_coh): each with its source; the composer does not rule who the evaluators were (open block)." }, { "act": "holding together (the archive's reading of the figure)", "sentences": [ 13, 15, 16 ], "claims": [ "F36", "D1443-01", "D1443-03", "D1443-04", "O1406-01", "O1406-02" ], "note": "the paper's 'synergistic effect' beside #1443's reading that the gain comes from the interface, ranked hypotheses not a verdict, and #1406's category with its falsifier: an archive mechanism on the field's own figure (ρ = extends); not composed as the paper's reading." }, { "act": "holding together (the dataset and the tradition)", "sentences": [ 9, 19 ], "claims": [ "F22", "D1443-02", "F48", "F49" ], "note": "the paper's statement that PHI synthesizes generations of scholarship and that its scale can correct the tradition it rests on, beside #1443's 'the dataset construction is the hidden achievement'." }, { "act": "holding together (the census against its sources)", "sentences": [ 8, 21 ], "claims": [ "RF-13", "RF-14", "F17", "F19", "RF-23", "F54", "F55" ], "note": "the census's 'not stated in the abstract' (period, count) and its licence 'not read' composed beside the paper's own statements of the period, the count and the licences, each with its locus; the census's fields are true of what it read and are reported as out of date against its saved source (notes)." }, { "act": "holding together (the homonym)", "sentences": [ 4, 5 ], "claims": [ "F8", "F9", "F10", "F62", "CAP-1-09", "CAP-1-12", "RF-02" ], "note": "the name composed as the makers give it, Homer's island, beside the encyclopedia's disambiguation and the registry's reading that the disambiguator selects the model; the island, the city and the other homonyms kept distinct entities, never senses of the name." } ], "plan": { "order": [ "F1", "RF-01", "RF-05", "RF-06", "RF-07", "RF-08", "RF-09", "RF-10", "F2", "F3", "F4", "F7", "F5", "F6", "F8", "F9", "F10", "RF-02", "F62", "CAP-1-09", "CAP-1-12", "RF-04", "F11", "F12", "F13", "RF-25", "RF-24", "F14", "F15", "F16", "F17", "F18", "F19", "F20", "F21", "RF-12", "RF-13", "RF-14", "F22", "D1443-02", "F23", "F24", "F25", "F26", "F27", "F28", "F29", "F30", "F31", "F32", "F33", "F34", "F35", "RF-16", "F36", "F38", "RF-15", "RF-11", "D1443-01", "F37", "F39", "F40", "F41", "D1443-03", "O1406-01", "O1406-02", "D1443-04", "D1443-05", "F42", "F43", "F44", "F45", "RF-17", "F46", "F47", "F48", "F49", "D1443-06", "F51", "RF-03", "F50", "F52", "F53", "RF-22", "RF-30", "F54", "F55", "RF-23", "F56", "F57", "F58", "F59", "F60", "RF-28", "RF-31", "RF-18", "RF-19", "RF-20", "RF-21", "RF-26", "RF-27", "RF-32", "F61", "RF-33", "RF-34", "RF-35", "CAP-1-06", "CAP-1-01", "CAP-1-11", "CAP-1-17", "RF-29", "CAP-1-02", "CAP-1-03", "CAP-1-05", "CAP-1-15", "CAP-1-13", "CAP-1-04", "CAP-1-14", "CAP-1-16", "CAP-1-10", "CAP-1-07", "CAP-1-08" ], "senses": { "what it is": [ 1, 2, 3 ], "the name": [ 4, 5 ], "the makers": [ 6 ], "the problem": [ 7 ], "the data": [ 8, 9 ], "the architecture": [ 10 ], "the outputs": [ 11 ], "the figures": [ 12 ], "the collaboration figure": [ 13, 14, 15 ], "the single-verdict hazard": [ 16 ], "redating Classical Athens": [ 17 ], "the stated limits": [ 18, 19 ], "access and licences": [ 20, 21 ], "genealogy and successors": [ 22, 23 ], "what the record asks": [ 24 ], "where it meets Revelation First": [ 25, 26 ], "the composition at the name": [ 27, 28, 29, 30 ] }, "case_by_sense": { "equal (one claim space; nothing translated)": [ "what it is", "the name", "the makers", "the problem", "the data", "the architecture", "the outputs", "the figures", "the collaboration figure", "the single-verdict hazard", "redating Classical Athens", "the stated limits", "access and licences", "genealogy and successors", "what the record asks", "where it meets Revelation First", "the composition at the name" ] }, "carried": { "archive": 60, "field": 62, "not_composed": {} }, "conflict_log": [ { "rule": "D_pres vs channel", "result": "no cut: the KO carries every field and archive claim; the compression's bound met with no channel log" }, { "rule": "P_coh", "result": "F37 (two evaluators 'with epigraphical expertise') and F38/F40 ('expert historians') beside F39 (student annotators and baseline evaluators) and F41 (Dickey: 'only made of students'), sentences 13–14; RF-13/RF-14/RF-23 (the census: period, count 'not stated', licence 'not read') beside F17, F19, F54, F55 (the paper states them), sentences 8 and 21; CAP-1-08 ('within thirty years and no closer') beside F34 (a median distance of 3 years), sentences 12 and 30; each composed with its source" }, { "rule": "N_c", "result": "no claim lowered by standing; D1443-03/D1443-04 at the survey's own grade (interpretation), O1406-01 as the definition it is and O1406-02 as its falsifier; F41 as contested, F40 as attributed; the paper's 'first' (F3, F58) as self-description" }, { "rule": "C_ex", "result": "senses composed ⊇ senses in the ledgers: every ledger lineage cited in the KO" } ] } }, "P": { "title": "Ithaca (DeepMind)", "lede": { "text": "is DeepMind's deep neural network, built with Ca' Foscari, Oxford and Athens partners, that restores, places and dates damaged ancient Greek inscriptions (Nature, 2022): inscriptions, not manuscripts, assistive by design (census record).", "claims": [ "F1", "F11", "F13", "RF-04", "RF-05", "RF-07", "RF-26", "F45" ] }, "sections": [ { "icon": "🧩", "head": "What it does", "items": [ { "label": "Restoration:", "text": "the top 20 hypotheses ranked by probability, with saliency maps (Nature; DeepMind).", "claims": [ "F27", "F28", "F31" ] }, { "label": "Place and date:", "text": "distributions over 84 regions and 160 decades (Nature; DeepMind).", "claims": [ "F29", "F30", "F18" ] }, { "label": "Build:", "text": "characters and words jointly; sparse-attention transformer blocks; three heads (Nature).", "claims": [ "F23", "F24", "F25", "F26" ] }, { "label": "Data:", "text": "PHI's inscriptions, 7th c. BC–5th c. AD, as I.PHI, 78,608 texts (Nature; census record); its construction \"the hidden achievement\" (#1443).", "claims": [ "F17", "F19", "F20", "F21", "F22", "RF-10", "RF-12", "RF-13", "RF-14", "D1443-02" ] } ] }, { "icon": "📊", "head": "What it scores", "items": [ { "label": "Alone:", "text": "62% restoration, 71% place, under 30 years from ground-truth ranges (Nature; census record).", "claims": [ "F32", "F33", "F34", "F35", "RF-16" ] }, { "label": "With historians:", "text": "25% to 72% (Nature; census record).", "claims": [ "F36", "F38", "RF-15", "RF-11" ] }, { "label": "Contested:", "text": "the evaluators were students, says Eleanor Dickey; the paper names two experts (Wikipedia; Nature).", "claims": [ "F41", "F40", "F37", "F39" ] }, { "label": "Athens:", "text": "held-out decrees dated to 421 BC (Nature).", "claims": [ "F42", "F43", "F44" ] } ] }, { "icon": "🔍", "head": "The archive's reading", "items": [ { "label": "The gain:", "text": "interface design, ranked hypotheses not a verdict (#1443).", "claims": [ "D1443-03", "D1443-01" ] }, { "label": "The hazard:", "text": "register suppression belongs to single-verdict interfaces, not machine reading as such (#1406; #1443).", "claims": [ "O1406-01", "O1406-02", "D1443-04", "D1443-05" ] }, { "label": "Unaskable:", "text": "its authors cannot be consulted (#1443).", "claims": [ "D1443-06" ] } ] }, { "icon": "⚠️", "head": "Its limits", "items": [ { "label": "Assistive:", "text": "designed to assist; restorations are plausible suggestions (Nature; census record).", "claims": [ "F45", "F46", "F47", "RF-17" ] }, { "label": "Tradition:", "text": "its data rests on the scholarship it corrects (Nature).", "claims": [ "F48", "F49" ] }, { "label": "Not the New Testament:", "text": "it dates inscriptions, not manuscripts or literary works (census record).", "claims": [ "RF-18", "RF-19", "RF-20", "RF-27", "RF-34", "RF-35" ] } ] }, { "icon": "📖", "head": "Revelation First", "items": [ { "label": "Reading:", "text": "a method to watch as machine dating nears papyri (census record); its methods claimed for papyrology (Nature).", "claims": [ "RF-21", "RF-32", "RF-33", "F61" ] } ] }, { "icon": "🔗", "head": "Access and lineage", "items": [ { "label": "Interface:", "text": "Predicting the Past, beside Aeneas (Predicting the Past; census record).", "claims": [ "F5", "F50", "F51", "RF-03", "RF-22", "RF-28" ] }, { "label": "Licences:", "text": "code Apache 2.0, weights CC BY-SA, article CC BY; the census's licence unread (Nature; census record).", "claims": [ "F54", "F55", "RF-23" ] }, { "label": "Lineage:", "text": "after Pythia (2019); other languages foreseen (Nature; DeepMind).", "claims": [ "F56", "F57", "F58", "F59", "F60" ] }, { "label": "Name:", "text": "Homer's island; bare, a homonym (Nature; Wikipedia; Capture Registry).", "claims": [ "F8", "F9", "F10", "F62", "RF-01", "RF-02", "CAP-1-09", "CAP-1-12" ] } ] }, { "icon": "🔎", "head": "At the name", "items": [ { "label": "Composed:", "text": "the model held; frame, year, Athens and partners absent; 'true historical creation date' (Capture Registry, 2026-10-31).", "claims": [ "CAP-1-01", "CAP-1-02", "CAP-1-03", "CAP-1-05", "CAP-1-06", "CAP-1-13", "CAP-1-15" ] }, { "label": "Cards:", "text": "regions, build, data and licence from GitHub and Nature; the weights' licence dropped; no card cites the record (Capture Registry).", "claims": [ "CAP-1-04", "CAP-1-14", "CAP-1-16", "CAP-1-10" ] }, { "label": "Analysis:", "text": "\"composed by its benchmark\"; thirty years, Enoch's margin (Capture Registry).", "claims": [ "CAP-1-07", "CAP-1-08" ] } ] } ], "open": [ { "question": "Did the collaboration figure measure qualified epigraphers or students?", "rests_on": [ "F36", "F37", "F38", "F39", "F40", "F41" ], "modality": "documented (Nature; the DeepMind post), attributed and contested (Wikipedia)", "would_resolve": "an account by the authors of the evaluators' qualifications, or an evaluation with qualified editors" }, { "question": "Does the 25-to-72 gain come from the ranked interface rather than the model's accuracy?", "rests_on": [ "D1443-03", "D1443-04", "O1406-01", "O1406-02", "F27", "F29", "F36" ], "modality": "interpretation (#1443), stipulation with its falsifier (#1406), documented (Nature)", "would_resolve": "historians evaluated with a single-verdict interface against the ranked one, at the same model" }, { "question": "How much of 'within 30 years' is a property of the dataset's own labels?", "rests_on": [ "F32", "F34", "F42", "F48", "F49", "RF-34", "CAP-1-08", "CAP-1-13" ], "modality": "documented and self-description (Nature), interpretation (census record; Capture Registry)", "would_resolve": "the distribution of dating errors by period and region, set against independent dates" }, { "question": "Will machine dating reach manuscripts, papyri or literary works, and the New Testament?", "rests_on": [ "F61", "F59", "F60", "RF-19", "RF-20", "RF-21", "RF-35" ], "modality": "interpretation (Nature; census record), documented (the DeepMind post; Wikipedia)", "would_resolve": "a model calibrated on a dated corpus of manuscripts or papyri, with its errors stated" }, { "question": "Which licences govern the interface and the repository as they now stand?", "rests_on": [ "RF-23", "RF-30", "F54", "F55", "F50" ], "modality": "documented (Nature; census record)", "would_resolve": "reading the repository and the interface's terms (GitHub unreachable from this environment; the interface a JavaScript application)" } ], "body_words": 319, "genre": "the compression: AIO's interaction grammar (lede, headed clusters, bolded labels, card rail), composed from the v0.9 plan in one claim space; it faces on the card, with T a record within it", "status": "composed 2026-10-31 from the v0.9 knowledge object; not frozen", "rail": [ { "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "claims": [ "F1", "RF-01", "RF-05", "F2", "F3", "F4", "F7", "F5", "F6" ], "earliest": "B4" }, { "lineage": "what the record asks: what evidence it holds, on which texts, for which questions", "claims": [ "RF-06", "RF-07", "RF-08", "RF-09", "RF-10", "RF-31", "RF-18", "RF-19", "RF-20" ], "earliest": "RF" }, { "lineage": "the name: Homer's island", "claims": [ "F8", "F9", "F10", "RF-02", "CAP-1-09", "CAP-1-12" ], "earliest": "B4" }, { "lineage": "reception: the encyclopedia and the bare name", "claims": [ "F62" ], "earliest": "B5" }, { "lineage": "the makers: DeepMind, its academic partners and the paper's authors", "claims": [ "RF-04", "F11", "F12", "F13", "RF-25", "RF-24" ], "earliest": "B4" }, { "lineage": "the problem: damaged, moved and undatable inscriptions", "claims": [ "F14", "F15", "F16" ], "earliest": "B4" }, { "lineage": "the data: PHI and I.PHI", "claims": [ "F17", "F18", "F19", "F20", "F21", "RF-12", "RF-13", "RF-14", "F22", "D1443-02" ], "earliest": "B4" }, { "lineage": "the architecture: characters and words, sparse attention, three heads", "claims": [ "F23", "F24", "F25", "F26" ], "earliest": "B4" }, { "lineage": "the outputs: ranked hypotheses, distributions, saliency", "claims": [ "F27", "F28", "F29", "F30", "F31" ], "earliest": "B4" }, { "lineage": "the headline figures: 62%, 71%, within 30 years", "claims": [ "F32", "F33", "F34", "F35", "RF-16" ], "earliest": "B4" }, { "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "claims": [ "F36", "F38", "RF-15", "RF-11", "D1443-01", "F37", "F39", "F40", "F41" ], "earliest": "B4" }, { "lineage": "the archive's reading: the gain from interface design, the single-verdict hazard", "claims": [ "D1443-03", "O1406-01", "O1406-02", "D1443-04", "D1443-05" ], "earliest": 1406 }, { "lineage": "redating Classical Athens", "claims": [ "F42", "F43", "F44" ], "earliest": "B4" }, { "lineage": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "claims": [ "F45", "RF-17", "F46", "F47", "F48", "F49" ], "earliest": "B4" }, { "lineage": "what the instrument cannot ask: authors who cannot be consulted", "claims": [ "D1443-06" ], "earliest": 1443 }, { "lineage": "access and licences: the interface, the code, the weights, the article", "claims": [ "F51", "RF-03", "F50", "F52", "F53", "RF-22", "RF-30", "F54", "F55", "RF-23" ], "earliest": "B4" }, { "lineage": "genealogy and successors: Pythia before, other languages and Aeneas after", "claims": [ "F56", "F57", "F58", "F59", "F60", "RF-28", "CAP-1-16" ], "earliest": "B4" }, { "lineage": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "claims": [ "RF-21", "RF-26", "RF-27", "RF-32", "RF-33", "RF-34", "RF-35", "CAP-1-08" ], "earliest": "RF" }, { "lineage": "scope: other ancient texts and disciplines", "claims": [ "F61" ], "earliest": "B4" }, { "lineage": "the composition at the name", "claims": [ "CAP-1-06", "CAP-1-01", "CAP-1-11", "CAP-1-17", "RF-29", "CAP-1-02", "CAP-1-03", "CAP-1-05", "CAP-1-15", "CAP-1-13", "CAP-1-04", "CAP-1-14", "CAP-1-10", "CAP-1-07" ], "earliest": "RF" } ], "entity": "Composed under v0.9, equal case (ρ = extends): KO_B sentences surviving 0 of 14; every field claim carried; the homonyms (Homer's island, the island of Ithaca, the city) kept distinct entities." }, "P_B": { "title": "Ithaca (the field alone)", "lede": { "text": "is a deep neural network by DeepMind and academic partners, presented in Nature in 2022, that restores, places and dates damaged ancient Greek inscriptions, designed to assist historians (Nature; the DeepMind post).", "claims": [ "F1", "F11", "F12", "F13", "F2", "F45" ] }, "sections": [ { "icon": "🧩", "head": "How it works", "items": [ { "label": "Data:", "text": "PHI's Greek inscriptions, processed into I.PHI, 78,608 texts (Nature).", "claims": [ "F17", "F19", "F20", "F21" ] }, { "label": "Outputs:", "text": "ranked restorations, region and date distributions, saliency (Nature; DeepMind).", "claims": [ "F27", "F29", "F30", "F31" ] } ] }, { "icon": "📊", "head": "Results", "items": [ { "label": "Alone:", "text": "62% restoration, 71% place, under 30 years (Nature; DeepMind).", "claims": [ "F32", "F33", "F34" ] }, { "label": "With historians:", "text": "25% to 72%; the evaluators contested as students (Nature; Wikipedia).", "claims": [ "F36", "F41" ] }, { "label": "Athens:", "text": "decrees redated to 421 BC (Nature).", "claims": [ "F42", "F43" ] } ] }, { "icon": "⚠️", "head": "Limits", "items": [ { "label": "Assistive:", "text": "plausible suggestions; data from the tradition (Nature).", "claims": [ "F46", "F47", "F48" ] } ] }, { "icon": "🔗", "head": "Access", "items": [ { "label": "Interface:", "text": "Predicting the Past, with Aeneas (Predicting the Past).", "claims": [ "F5", "F50" ] }, { "label": "Licences:", "text": "Apache 2.0 code; CC BY-SA weights (Nature).", "claims": [ "F54" ] }, { "label": "Name:", "text": "Homer's island; no Wikipedia article (Wikipedia).", "claims": [ "F9", "F62" ] } ] } ], "open": [ { "question": "Did the collaboration figure measure qualified epigraphers?", "rests_on": [ "F37", "F39", "F41" ], "modality": "documented (Nature), contested (Wikipedia)", "would_resolve": "an account of the evaluators' qualifications" }, { "question": "Will the methods reach other ancient texts?", "rests_on": [ "F59", "F61" ], "modality": "documented and interpretation (the DeepMind post; Nature)", "would_resolve": "a published model for papyri or manuscripts" } ], "body_words": 116, "arm": "B (the field alone)", "genre": "popup — the compression from the field ledger alone (the four cards and Wikipedia)", "status": "composed 2026-10-31 from the field ledger alone; not frozen", "rail": [ { "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "claims": [ "F1", "F2", "F3", "F4", "F5", "F6", "F7" ], "earliest": "B4" }, { "lineage": "the makers: DeepMind, its academic partners and the paper's authors", "claims": [ "F12", "F13", "F11" ], "earliest": "B4" }, { "lineage": "the name: Homer's island", "claims": [ "F8", "F9", "F10" ], "earliest": "B4" }, { "lineage": "the problem: damaged, moved and undatable inscriptions", "claims": [ "F14", "F15", "F16" ], "earliest": "B4" }, { "lineage": "the data: PHI and I.PHI", "claims": [ "F17", "F18", "F19", "F20", "F21", "F22" ], "earliest": "B4" }, { "lineage": "the architecture: characters and words, sparse attention, three heads", "claims": [ "F23", "F24", "F25", "F26" ], "earliest": "B4" }, { "lineage": "the outputs: ranked hypotheses, distributions, saliency", "claims": [ "F27", "F28", "F29", "F30", "F31" ], "earliest": "B4" }, { "lineage": "the headline figures: 62%, 71%, within 30 years", "claims": [ "F32", "F33", "F34", "F35" ], "earliest": "B4" }, { "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "claims": [ "F36", "F38", "F37", "F39", "F40", "F41" ], "earliest": "B4" }, { "lineage": "redating Classical Athens", "claims": [ "F42", "F43", "F44" ], "earliest": "B4" }, { "lineage": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "claims": [ "F45", "F46", "F47", "F48", "F49" ], "earliest": "B4" }, { "lineage": "access and licences: the interface, the code, the weights, the article", "claims": [ "F50", "F51", "F52", "F53", "F54", "F55" ], "earliest": "B4" }, { "lineage": "genealogy and successors: Pythia before, other languages and Aeneas after", "claims": [ "F56", "F57", "F58", "F59", "F60" ], "earliest": "B4" }, { "lineage": "scope: other ancient texts and disciplines", "claims": [ "F61" ], "earliest": "B4" }, { "lineage": "reception: the encyclopedia and the bare name", "claims": [ "F62" ], "earliest": "B5" } ], "entity": "With no archive claim admitted, the entry is the model at its paper's and its makers' resolution, with the encyclopedia's contest of the evaluators." }, "KO_B": { "title": "Ithaca (the field alone)", "arm": "B (the field alone)", "genre": "knowledge object", "status": "composed 2026-10-31; not frozen", "sentences": [ { "n": 1, "para": 0, "text": "Ithaca is a deep neural network \"for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions\", presented in Nature on 9 March 2022 by Yannis Assael, Thea Sommerschield and colleagues of DeepMind, Ca' Foscari, Oxford and the Athens University of Economics and Business (Nature; the DeepMind post).", "claims": [ "F1", "F12", "F13", "F11" ] }, { "n": 2, "para": 0, "text": "The post states the work as restoring, placing and dating ancient texts through collaboration between AI and historians, and the repository and interface site as a model by DeepMind and academic partners that assists historians with Greek and Latin inscriptions beside Aeneas (the DeepMind post; GitHub card; Predicting the Past).", "claims": [ "F2", "F3", "F4", "F5", "F6", "F7" ] }, { "n": 3, "para": 0, "text": "It is named after the Greek island in Homer's Odyssey (Nature; the DeepMind post; Wikipedia).", "claims": [ "F8", "F9", "F10" ] }, { "n": 4, "para": 1, "text": "Inscriptions are damaged, moved and undatable by radiocarbon (Nature; the DeepMind post).", "claims": [ "F14", "F15", "F16" ] }, { "n": 5, "para": 1, "text": "It was trained on PHI, 178,551 Greek inscriptions of the seventh century BC to the fifth century AD in 84 regions, processed into I.PHI of 78,608, a dataset that synthesizes generations of scholarship (Nature; the DeepMind post).", "claims": [ "F17", "F18", "F19", "F20", "F21", "F22" ] }, { "n": 6, "para": 2, "text": "It reads characters and words jointly through eight sparse-attention transformer blocks and three task heads (Nature; the DeepMind post).", "claims": [ "F23", "F24", "F25", "F26" ] }, { "n": 7, "para": 2, "text": "It returns the top 20 ranked restorations, distributions over regions and over 160 decades, and saliency maps (Nature; the DeepMind post).", "claims": [ "F27", "F28", "F29", "F30", "F31" ] }, { "n": 8, "para": 3, "text": "It restores at 62% and attributes place at 71%, dating to less than 30 years of the ground-truth ranges, a median of 3 years (Nature; the DeepMind post; Wikipedia).", "claims": [ "F32", "F33", "F34", "F35" ] }, { "n": 9, "para": 3, "text": "Historians using it rose from 25% to 72% (Nature; the DeepMind post); the baseline was two evaluators with epigraphical expertise, the paper thanks student evaluators, and Eleanor Dickey noted the test was only of students (Nature; Wikipedia).", "claims": [ "F36", "F38", "F37", "F39", "F40", "F41" ] }, { "n": 10, "para": 3, "text": "Held-out Athenian decrees are dated to 421 BC on average, 5 years from the modern lower dating (Nature; the DeepMind post).", "claims": [ "F42", "F43", "F44" ] }, { "n": 11, "para": 4, "text": "It is designed to assist and expand the historian's workflow, its restorations plausible suggestions, its data resting on the tradition it can correct, its discretized dates a noted limit (Nature).", "claims": [ "F45", "F46", "F47", "F48", "F49" ] }, { "n": 12, "para": 5, "text": "It is used through Predicting the Past, the code under Apache 2.0, the weights under CC BY-SA 4.0 and the article under CC BY 4.0 (Predicting the Past; Nature; the DeepMind post).", "claims": [ "F50", "F51", "F52", "F53", "F54", "F55" ] }, { "n": 13, "para": 5, "text": "It follows Pythia (2019), is called the first model of its kind, and versions for other ancient languages and disciplines are foreseen (Nature; the DeepMind post; Wikipedia).", "claims": [ "F56", "F57", "F58", "F59", "F60", "F61" ] }, { "n": 14, "para": 6, "text": "Wikipedia has no article on it; the bare name is a disambiguation page for Homer's island and others (Wikipedia).", "claims": [ "F62" ] } ] }, "L_B": { "words": 399, "text": "Ithaca is a deep neural network \"for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions\", presented in Nature on 9 March 2022 by Yannis Assael, Thea Sommerschield and colleagues of DeepMind, Ca' Foscari, Oxford and the Athens University of Economics and Business (Nature; the DeepMind post). The post states the work as restoring, placing and dating ancient texts through collaboration between AI and historians, and the repository and interface site as a model by DeepMind and academic partners that assists historians with Greek and Latin inscriptions beside Aeneas (the DeepMind post; GitHub card; Predicting the Past). It is named after the Greek island in Homer's Odyssey (Nature; the DeepMind post; Wikipedia). Inscriptions are damaged, moved and undatable by radiocarbon (Nature; the DeepMind post). It was trained on PHI, 178,551 Greek inscriptions of the seventh century BC to the fifth century AD in 84 regions, processed into I.PHI of 78,608, a dataset that synthesizes generations of scholarship (Nature; the DeepMind post). It reads characters and words jointly through eight sparse-attention transformer blocks and three task heads (Nature; the DeepMind post). It returns the top 20 ranked restorations, distributions over regions and over 160 decades, and saliency maps (Nature; the DeepMind post). It restores at 62% and attributes place at 71%, dating to less than 30 years of the ground-truth ranges, a median of 3 years (Nature; the DeepMind post; Wikipedia). Historians using it rose from 25% to 72% (Nature; the DeepMind post); the baseline was two evaluators with epigraphical expertise, the paper thanks student evaluators, and Eleanor Dickey noted the test was only of students (Nature; Wikipedia). Held-out Athenian decrees are dated to 421 BC on average, 5 years from the modern lower dating (Nature; the DeepMind post). It is designed to assist and expand the historian's workflow, its restorations plausible suggestions, its data resting on the tradition it can correct, its discretized dates a noted limit (Nature). It is used through Predicting the Past, the code under Apache 2.0, the weights under CC BY-SA 4.0 and the article under CC BY 4.0 (Predicting the Past; Nature; the DeepMind post). It follows Pythia (2019), is called the first model of its kind, and versions for other ancient languages and disciplines are foreseen (Nature; the DeepMind post; Wikipedia). Wikipedia has no article on it; the bare name is a disambiguation page for Homer's island and others (Wikipedia).", "note": "L_F(B), the field composed in the field's ontology (v0.9 §0.4), realized as KO_B; its rail is P_B's, by lineage." }, "L_BA": { "words": 2021, "text": "Ithaca is a deep neural network \"for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions\" (Nature, 2022), which the Revelation First census record names Ithaca and files under AI and ancient texts, \"machine reading, restoration and dating\", as evidence of the kind \"machine reading, restoration or dating\" whose unit is \"an inscription's restored text, attributed place and attributed date\" and whose texts are ancient Greek inscriptions (census record). Its makers state the work as \"Restoring, placing, and dating ancient texts through collaboration between AI and historians\" and introduce Ithaca as the first deep neural network to restore the missing text of damaged inscriptions, identify their original location and help establish their date (the DeepMind post); the repository's card gives \"a deep neural network developed by DeepMind and academic partners\" (GitHub card), and the encyclopedia one that \"helps researchers restore the empty text of damaged Greek documents\" and identify their date and origin (Wikipedia). The interface site calls Ithaca and Aeneas \"generative AI models designed to assist historians in the study of ancient Greek and Latin inscriptions\", part of \"a wider initiative on developing collaborative AI for historical research\" (Predicting the Past). The name is the island's: Ithaca \"was named after the Greek island that eluded the hero Odysseus’ homecoming\" (Nature), after \"the Greek island in Homer’s Odyssey\" (the DeepMind post; Wikipedia), and the census gives its other name as \"Ithaca (DeepMind)\" (census record). On the encyclopedia the bare name \"most commonly refers to\" Homer's Ithaca and is a disambiguation page that does not list the model, which has no article of its own (Wikipedia); the Capture Registry records that the operator \"added the disambiguator to a name that is bare a homonym\", and reads that 'DeepMind' selects the model from the island and the city (Capture Registry, 2026-10-31). Its operator is \"DeepMind with historians and collaborators; described in Nature (2022)\" (census record): DeepMind worked with the Department of Humanities of Ca' Foscari University of Venice, the Classics Faculty of the University of Oxford and the Department of Informatics of the Athens University of Economics and Business (the DeepMind post), the paper's co-first authors are Yannis Assael and Thea Sommerschield with those affiliations and DeepMind in London, and the paper was published in Nature 603, 280–283, on 9 March 2022 (Nature), the census dating it to 2022 and its lifecycle \"published 2022; interface online\" (census record). The paper's problem is that inscriptions \"have been damaged to the point of illegibility, transported far from their original location and their date of writing is steeped in uncertainty\", and that \"radiocarbon dating is unusable owing to the inorganic nature of most inscribed supports\" (Nature), which the post gives as dating techniques that \"cannot be used on these materials\" (the DeepMind post). It was trained on Greek inscriptions from across the ancient Mediterranean between the seventh century BC and the fifth century AD, from the Packard Humanities Institute's Searchable Greek Inscriptions (PHI), 178,551 transcribed texts in 84 ancient regions, processed into I.PHI, \"containing 78,608 inscriptions\" (Nature; the DeepMind post); the census gives its geography as the ancient Greek world and, from the abstract, its period as \"not stated in the abstract\" and its count as \"not stated in the saved abstract\" (census record). The paper states that PHI \"is a synthesis of generations of scholarly research\" (Nature), and the archive's field survey reads the processing as the work itself: the model assumes \"A large, clean, machine-actionable corpus\", I.PHI \"required an extended ruleset to normalise 84 regions and heterogeneous date formats\", and \"The dataset construction is the hidden achievement\" (#1443). The text is processed \"as character and word representations jointly\" through eight stacked transformer decoder blocks after BigBird, each with four sparse attention heads, and three task heads for restoration, place and date (Nature), the post describing a sparse self-attention core that evaluates words and characters in parallel (the DeepMind post). For restoration it offers \"the top 20 decoded predictions ranked by probability\" (Nature), several hypotheses \"for historians to choose from using their expertise\"; for place it shows its uncertainty as a probability distribution \"instead of just a single output\"; for date it predicts over 160 decades from 800 BC to AD 800; and saliency maps show which inputs drove a prediction (the DeepMind post; Nature). Alone it restores damaged texts at 62% and attributes place at 71%, and it dates \"to less than 30 years of their ground-truth ranges\" (Nature; the DeepMind post; Wikipedia's 'dating precision of 30 years'); measured, that is a 26.3% character error rate and 61.8% top-1 restoration, 70.8% top-1 and 82.1% top-3 for regions, and an average distance of 29.3 years, \"with a median distance of only 3 years\", from the target brackets (Nature), which the census carries as its second finding (census record). With it, historians' restoration accuracy rose \"from 25% to 72%\", which the paper reads as \"the synergistic effect of this research tool\" (Nature) and the post as expert historians \"surpassing the model’s individual performance\" (the DeepMind post); the census carries it as its first finding and the method as \"evaluated alone and with historians\" (census record), and the archive's survey restates the figures from the paper (#1443). The paper's human baseline is \"two evaluators with epigraphical expertise (‘ancient historian’)\", and it thanks \"the student annotators and baseline evaluators\" of the Athens University of Economics and Business's MSc in Digital Methods for the Humanities and of Oxford's Faculty of Classics (Nature); the encyclopedia reports the authors' claim for \"expert historians\" and that \"Eleanor Dickey noted that this test was actually only made of students\", it being unclear how helpful Ithaca would be to \"genuinely qualified editors\" (Wikipedia). The archive's field survey reads the figure as a property of the interface: \"this is not \"AI helps.\"\" — \"The 25→72 gain cannot come from accuracy: the model alone scores 62, below the collaborative figure\"; it comes from interface design, \"ranked top-20 hypotheses with saliency, not a verdict; the historian evaluates alternatives rather than accepting an answer\" (#1443). The archive defines differential register prioritization as \"a condition in which multiple interpretive fields are available, but some are weighted in ways that suppress or erase others, materially altering the reading of ambiguous language\", a specification weakened \"if reading systems can be shown to reliably recover suppressed registers without branch-preservation\" (#1406); its field survey reads Ithaca as \"empirical evidence that the hazard is a property of single-verdict interfaces, not of machine reading as such\", with ranked-alternative presentation the mitigation, which \"converts an internal warning into a design requirement with external validation\" (#1443). On the disputed Athenian decrees, held out of training, its predictions average 421 BC and sit on average 5 years from the modern 'lower' dating where the dataset's labels sit 27 years off (Nature), the work by which the paper says it is \"redating key texts of Classical Athens\" (Nature) and historians \"have already used the tool to reevaluate significant periods in Greek history\" (the DeepMind post). Its frame is stated in the abstract, \"Ithaca is designed to assist and expand the historian’s workflow\", its architecture focused on \"collaboration, decision support and interpretability\" (Nature; census record): \"an assistive tool\" that proposes \"possible solutions for the scholar to consider\", none of whose restorations \"are assumed to be objectively certain—instead, they are presented as plausible suggestions\" (Nature). The paper holds that the scale of its data can challenge errors in the academic tradition \"notwithstanding the fact that the dataset in question is originally based on the accumulated academic tradition\", and that \"The limitations of discretizing and amalgamating date ranges of different levels of precision based on past scholarship have been noted\", which its scale at least partially meets (Nature); the archive's survey states what such an instrument cannot ask: attestation is answerable \"only where there is a living author who can sign, date, photograph, enroll, attest, consent, and be asked again next year\", and \"Vesuvius and Ithaca work on authors who cannot be consulted\" (#1443). It is used through \"an open-source and publicly available interface\" at ithaca.deepmind.com (Nature; census record), now served as Predicting the Past, \"Contextualising, restoring, and attributing ancient texts\" (Predicting the Past), launched with Google Cloud and Google Arts & Culture with the code, the pretrained model and a Colaboratory notebook open-sourced, its web interface built by Justin Grayston, Benjamin Maynard and Ricardo Cardenas of Google Cloud (the DeepMind post); the census reads the saved page as \"a JavaScript application\" with the paper open in Nature and left \"the interface (JavaScript)\" and \"the training corpus\" unchecked (census record). The training and inference code is on GitHub \"under Apache License 2.0, along with the trained weights, licensed under Creative Commons Attribution-ShareAlike 4.0 International\", and the article is licensed under CC BY 4.0 (Nature); the census records its licence as \"not read\" (census record). Its closest predecessor is Pythia, the authors' 2019 restoration model, which it builds on and extends (Nature; the DeepMind post; Wikipedia), and the paper calls it \"to our knowledge the first epigraphic restoration and attribution model of its kind\" (Nature). Versions for other ancient languages were in progress, \"from Akkadian to Demotic and Hebrew to Mayan\" (the DeepMind post; Wikipedia); the census names its neighbour Aeneas as \"DeepMind's later model for Latin inscriptions\" (census record), the model the interface site presents beside it. The census asks of each resource \"What evidence does each resource hold on when and from what the New Testament writings were composed, and how do machines now read those texts?\", and answers for Ithaca \"how accurately a model restores, places and dates Greek inscriptions, and how much it helps historians\", not \"the dating of manuscripts or literary works\" nor \"anything on the New Testament\" (census record). Its reading, the archive's: \"Machine dating of Greek text: inscriptions, not manuscripts. A method the dating workstream can watch; it bears on the thesis only as machine dating moves toward literary and papyrus evidence\", with the tags \"NOT a manuscript-dating model\" and \"NOT a New Testament resource\", under the census's correction that \"Revelation First is the priority thesis\", not the pre-70 date (census record); the authors, for their part, hold the methods applicable \"to all disciplines dealing with ancient text (papyrology, numismatics, codicology), to any language (ancient or modern)\" (Nature). In the worked question of what machine dating of manuscripts can establish, Ithaca stands for \"restoration, place and date for Greek inscriptions, within 30 years of ground truth\": \"Machine dating is calibrated on a corpus. What it establishes holds for that corpus and that kind of evidence\", and \"None of the four dates a New Testament papyrus or a composition\" (census record), the thirty years being the paper's distance from \"ground-truth ranges\" (Nature). At 'Ithaca DeepMind' on 31 October 2026 Google's AI Overview gave \"the model, its three tasks with their figures, how it is built and where it can be used, from its maker, its code and its paper\": \"THE MODEL HELD\", the figures the record's own from the Nature abstract, \"Google DeepMind named; no person; no identifier\", at a dimensional PER of 0.105 (Capture Registry, 2026-10-31), the address carrying a disambiguator because the bare name is shared (census record). The paper's frame was \"carried only by the collaboration figures\", the redating of Classical Athens and the model's year absent, the \"academic partners\" unnamed, and \"within 30 years of their true historical creation date\" an inflation of the record's \"ground-truth ranges\", the paper's words (Nature), the plausible suggestions not carried (Capture Registry, 2026-10-31). The 84 regions, the architecture, the PHI dataset and the Apache-2.0 licence, each stated in the paper (Nature), came \"from the GitHub and Nature cards\", the record not having read the licence, the CC BY-SA weights not carried, the closing offer tying Ithaca to \"newer models like Aeneas\", and no card citing the census record, built the day before (Capture Registry, 2026-10-31). The registry's analysis: \"A model is composed by its benchmark, accurately\", and \"What the field takes from Ithaca is the scale of the error: a machine dates an inscription to within thirty years and no closer, which is the same margin as Enoch's for the scrolls\" (Capture Registry, 2026-10-31).", "rail": [], "note": "L(B ∪ A) in the equal-terms case (v0.9 §0.4), realized as the knowledge object (objects.KO); its rail is the compression's, by lineage." } }, "field": [ { "id": "B1", "card": "Google DeepMind — Predicting the past with Ithaca — Google DeepMind", "fetched": "card; B1.html sha256 832d3d6cfd97b40a, HTTP 200 at 2026-10-31T04:39:12Z", "origin": "card" }, { "id": "B2", "card": "GitHub — GitHub - google-deepmind/ithaca: Restoring and attributing ancient texts using deep neural networks · GitHub", "fetched": "card; unreachable: HTTP 403 from the build environment's GitHub gate at 2026-10-31T04:39:12Z (B2-403.html sha256 b3e03aaaff81d687); not retried; the card's snippet stands", "origin": "card" }, { "id": "B3", "card": "Predicting the Past — Predicting the Past", "fetched": "card; B3.html sha256 08a11cf87711fe87, HTTP 200 at 2026-10-31T04:39:13Z; byte-identical to the census copy ithaca--home; text from its title and meta description (B3-meta.txt)", "origin": "card" }, { "id": "B4", "card": "Nature — Restoring and attributing ancient texts using deep neural networks | Nature", "fetched": "card; B4.html sha256 7bab9e1b3786b154, HTTP 200 at 2026-10-31T04:39:15Z; text identical to the census copy ithaca--nature (sha256 d2b4df37a93252c3)", "origin": "card" }, { "id": "B5", "card": "Wikipedia — Google DeepMind §Others, 'Archaeology' (no article 'Ithaca (neural network)' or similar; 'Ithaca' a disambiguation page that does not list the model)", "fetched": "wikipedia; no article (raw wikitext HTTP 404 for five titles; 'Ithaca' a disambiguation page not listing the model); the section 'Google DeepMind' › 'Archaeology' fetched, B5.html sha256 4e7bfba0afa24409, HTTP 200 at 2026-10-31T04:41:24Z", "origin": "wikipedia" } ], "field_claims": { "F1": { "source": "B4", "origin": "card", "locus": "Abstract s3", "quote": "Here we present Ithaca, a deep neural network for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions.", "claim": "the paper's definition of its subject: Ithaca, a deep neural network for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F2": { "source": "B1", "origin": "card", "locus": "standfirst", "quote": "Restoring, placing, and dating ancient texts through collaboration between AI and historians", "claim": "the post's statement of the work: restoring, placing and dating ancient texts through collaboration between AI and historians", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F3": { "source": "B1", "origin": "card", "locus": "¶3 s1", "quote": "we jointly introduce Ithaca, the first deep neural network that can restore the missing text of damaged inscriptions, identify their original location, and help establish the date they were created.", "claim": "Ithaca introduced as the first deep neural network that restores the missing text of damaged inscriptions, identifies their original location and helps establish their date", "modality": "self-description", "kind": "definition", "kernel": false, "const": false, "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F4": { "source": "B2", "origin": "card", "locus": "card snippet (unreachable card)", "quote": "Overview: Ithaca is a deep neural network developed by DeepMind and academic partners for restoring, geographically attributing, and chronologically dating anci...", "claim": "the repository's overview as the card gives it: a deep neural network by DeepMind and academic partners for restoring, geographically attributing and chronologically dating ancient texts (snippet truncated)", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "date": "2026-10-31 (the card as rendered; the page unreachable)", "quote_ok": true, "checked_against": "the card's snippet (unreachable card), as in the registry's cite_list" }, "F5": { "source": "B3", "origin": "card", "locus": "meta description s1", "quote": "Ithaca and Aeneas are generative AI models designed to assist historians in the study of ancient Greek and Latin inscriptions.", "claim": "the interface site's account: Ithaca and Aeneas as generative AI models designed to assist historians in the study of ancient Greek and Latin inscriptions", "modality": "stipulation", "kind": "definition", "kernel": true, "const": false, "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "date": "2026-10-31 (read; the page is undated)", "quote_ok": true, "checked_against": "saved text B3-meta.txt (sha256 3f6895b72c46ded7…), whitespace-normalised" }, "F6": { "source": "B3", "origin": "card", "locus": "meta description s2", "quote": "They are part of a wider initiative on developing collaborative AI for historical research.", "claim": "the two models are part of a wider initiative on collaborative AI for historical research", "modality": "stipulation", "kind": "function", "kernel": false, "const": false, "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "date": "2026-10-31 (read; the page is undated)", "quote_ok": true, "checked_against": "saved text B3-meta.txt (sha256 3f6895b72c46ded7…), whitespace-normalised" }, "F7": { "source": "B5", "origin": "wikipedia", "locus": "Google DeepMind §Others, Archaeology s2", "quote": "This deep neural network helps researchers restore the empty text of damaged Greek documents, and to identify their date and geographical origin.", "claim": "the encyclopedia's account: a deep neural network that helps researchers restore the missing text of damaged Greek documents and identify their date and origin", "modality": "documented", "kind": "definition", "kernel": false, "const": false, "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "date": "2026-10-31 (read; Wikipedia revision 1380412362)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 e0065367800a53a0…), whitespace-normalised" }, "F8": { "source": "B4", "origin": "card", "locus": "Main, 'Deep learning for epigraphy' ¶2", "quote": "Ithaca, which was named after the Greek island that eluded the hero Odysseus’ homecoming", "claim": "the model is named after the Greek island that eluded Odysseus' homecoming", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the name: Homer's island", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F9": { "source": "B1", "origin": "card", "locus": "¶3 s2", "quote": "Ithaca is named after the Greek island in Homer’s Odyssey and builds upon and extends Pythia , our previous system that focused on textual restoration.", "claim": "named after the Greek island in Homer's Odyssey; builds on and extends Pythia, DeepMind's earlier system for textual restoration", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the name: Homer's island", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F10": { "source": "B5", "origin": "wikipedia", "locus": "Google DeepMind §Others, Archaeology s1", "quote": "Google has unveiled a new archaeology document program, named Ithaca after the Greek island in Homer's Odyssey .", "claim": "the encyclopedia: Google's archaeology document program, named after the Greek island in Homer's Odyssey", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "the name: Homer's island", "date": "2026-10-31 (read; Wikipedia revision 1380412362)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 e0065367800a53a0…), whitespace-normalised" }, "F11": { "source": "B1", "origin": "card", "locus": "¶2", "quote": "we collaborated with the Department of Humanities of Ca' Foscari University of Venice , the Classics Faculty of the University of Oxford , and the Department of Informatics of the Athens University of Economics and Business", "claim": "DeepMind's academic partners: Ca' Foscari University of Venice (Humanities), the University of Oxford (Classics), the Athens University of Economics and Business (Informatics)", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "the makers: DeepMind, its academic partners and the paper's authors", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F12": { "source": "B4", "origin": "card", "locus": "Author information, Author notes and Affiliations", "quote": "These authors contributed equally: Yannis Assael, Thea Sommerschield … DeepMind, London, UK … Department of Humanities, Ca’ Foscari University of Venice, Venice, Italy … Department of Informatics, Athens University of Economics and Business, Athens, Greece … Faculty of Classics, University of Oxford, Oxford, UK", "claim": "the paper's co-first authors, Yannis Assael and Thea Sommerschield, and its affiliations: DeepMind (London), Ca' Foscari, the Athens University of Economics and Business, the Oxford Faculty of Classics", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "the makers: DeepMind, its academic partners and the paper's authors", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised; multi-part quote, each part checked" }, "F13": { "source": "B4", "origin": "card", "locus": "article header", "quote": "Published: 09 March 2022 … Nature volume 603 , pages 280–283 ( 2022 )", "claim": "the paper: Nature 603, 280–283, published 9 March 2022", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the makers: DeepMind, its academic partners and the paper's authors", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised; multi-part quote, each part checked" }, "F14": { "source": "B4", "origin": "card", "locus": "Abstract s2", "quote": "many inscriptions have been damaged to the point of illegibility, transported far from their original location and their date of writing is steeped in uncertainty.", "claim": "the problem the paper addresses: inscriptions damaged to illegibility, moved from their original location, and uncertain in date", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the problem: damaged, moved and undatable inscriptions", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F15": { "source": "B4", "origin": "card", "locus": "Main ¶1", "quote": "radiocarbon dating is unusable owing to the inorganic nature of most inscribed supports", "claim": "radiocarbon dating cannot be used on most inscribed supports, which are inorganic", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the problem: damaged, moved and undatable inscriptions", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F16": { "source": "B1", "origin": "card", "locus": "¶1", "quote": "modern dating techniques, such as radiocarbon dating , cannot be used on these materials, making inscriptions difficult and time-consuming to interpret.", "claim": "radiocarbon dating cannot be used on stone, pottery and metal, which makes inscriptions slow to interpret", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the problem: damaged, moved and undatable inscriptions", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F17": { "source": "B4", "origin": "card", "locus": "Main, 'Deep learning for epigraphy' ¶2", "quote": "was trained on inscriptions written in the ancient Greek language and across the ancient Mediterranean world between the seventh century bc and the fifth century ad", "claim": "trained on Greek inscriptions from across the ancient Mediterranean, seventh century BC to fifth century AD", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the data: PHI and I.PHI", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F18": { "source": "B4", "origin": "card", "locus": "Main, 'Working with Greek inscriptions'", "quote": "dataset, which consists of the transcribed texts of 178,551 inscriptions … PHI lists a total of 84 ancient regions", "claim": "the PHI dataset holds 178,551 transcribed inscriptions and lists 84 ancient regions", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the data: PHI and I.PHI", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised; multi-part quote, each part checked" }, "F19": { "source": "B4", "origin": "card", "locus": "Main, 'Working with Greek inscriptions'", "quote": "the resulting dataset I.PHI is to our knowledge the largest multitask dataset of machine-actionable epigraphical text, containing 78,608 inscriptions.", "claim": "the processed dataset I.PHI, 78,608 inscriptions, the largest multitask dataset of machine-actionable epigraphical text to the authors' knowledge", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the data: PHI and I.PHI", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F20": { "source": "B4", "origin": "card", "locus": "Data availability", "quote": "Ithaca was trained on The Packard Humanities Institute’s Searchable Greek Inscriptions public dataset, PHI", "claim": "the training source: the Packard Humanities Institute's Searchable Greek Inscriptions public dataset (PHI)", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the data: PHI and I.PHI", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F21": { "source": "B1", "origin": "card", "locus": "'Collaborative tools' ¶1", "quote": "Ithaca is trained on the largest digital dataset of Greek inscriptions from the Packard Humanities Institute .", "claim": "trained on the largest digital dataset of Greek inscriptions, the Packard Humanities Institute's", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the data: PHI and I.PHI", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F22": { "source": "B4", "origin": "card", "locus": "Methods, 'Data circularity'", "quote": "Ithaca’s source dataset (PHI) is a synthesis of generations of scholarly research.", "claim": "the source dataset is a synthesis of generations of scholarly research", "modality": "self-description", "kind": "fact", "kernel": false, "const": false, "lineage": "the data: PHI and I.PHI", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F23": { "source": "B4", "origin": "card", "locus": "Main, 'Ithaca is a model for epigraphic tasks' ¶1", "quote": "we process the input text as character and word representations jointly", "claim": "the input is processed as character and word representations jointly", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the architecture: characters and words, sparse attention, three heads", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F24": { "source": "B4", "origin": "card", "locus": "Methods, 'Torso'", "quote": "The architecture of Ithaca’s torso consists of eight stacked transformer decoder blocks, inspired by the large-scale transformer model BigBird 73 . Every block uses four sparse attention heads", "claim": "the torso: eight stacked transformer decoder blocks after BigBird, each with four sparse attention heads", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the architecture: characters and words, sparse attention, three heads", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F25": { "source": "B4", "origin": "card", "locus": "Main, 'Ithaca is a model for epigraphic tasks' ¶2", "quote": "The final output of the torso is passed to three different task heads that handle restoration, geographical attribution and chronological attribution, respectively.", "claim": "three task heads, for restoration, geographical attribution and chronological attribution", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the architecture: characters and words, sparse attention, three heads", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F26": { "source": "B1", "origin": "card", "locus": "'Collaborative tools' ¶1", "quote": "we trained it using both words and the individual characters as inputs. The sparse self-attention mechanism at the model’s core evaluates these two inputs in parallel", "claim": "trained on words and individual characters as inputs, evaluated in parallel by a sparse self-attention mechanism", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the architecture: characters and words, sparse attention, three heads", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F27": { "source": "B4", "origin": "card", "locus": "Main, 'Interpreting the outputs' ¶1", "quote": "Ithaca offers a set of the top 20 decoded predictions ranked by probability", "claim": "for restoration the model offers the top 20 predictions ranked by probability, not a single hypothesis", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the outputs: ranked hypotheses, distributions, saliency", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F28": { "source": "B1", "origin": "card", "locus": "'Collaborative tools', visual aids (Restoration hypotheses)", "quote": "Ithaca generates several prediction hypotheses for the text restoration task for historians to choose from using their expertise.", "claim": "several restoration hypotheses for historians to choose from by their expertise", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the outputs: ranked hypotheses, distributions, saliency", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F29": { "source": "B1", "origin": "card", "locus": "'Collaborative tools', visual aids (Geographical attribution)", "quote": "Ithaca shows its uncertainty by giving historians a probability distribution over all possible predictions – instead of just a single output.", "claim": "the model shows its uncertainty as a probability distribution over all predictions instead of a single output", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the outputs: ranked hypotheses, distributions, saliency", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F30": { "source": "B4", "origin": "card", "locus": "Main, 'Interpreting the outputs' ¶3", "quote": "Ithaca discretizes all dates between 800 bc and ad 800 into 10-year bins, resulting in 160 decades.", "claim": "dates are predicted as a distribution over 160 decades from 800 BC to AD 800", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the outputs: ranked hypotheses, distributions, saliency", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F31": { "source": "B1", "origin": "card", "locus": "'Collaborative tools', visual aids (Saliency maps)", "quote": "Ithaca uses a technique commonly used in computer vision that identifies which input sequences contribute most to a prediction.", "claim": "saliency maps show which input sequences contribute most to a prediction", "modality": "documented", "kind": "function", "kernel": false, "const": false, "lineage": "the outputs: ranked hypotheses, distributions, saliency", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F32": { "source": "B4", "origin": "card", "locus": "Abstract s7", "quote": "Ithaca can attribute inscriptions to their original location with an accuracy of 71% and can date them to less than 30 years of their ground-truth ranges, redating key texts of Classical Athens and contributing to topical debates in ancient history.", "claim": "71% accuracy in attributing location; dates to less than 30 years of the ground-truth ranges; redates key texts of Classical Athens", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the headline figures: 62%, 71%, within 30 years", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F33": { "source": "B1", "origin": "card", "locus": "¶3 s3", "quote": "Our evaluations show that Ithaca achieves 62% accuracy in restoring damaged texts, 71% accuracy in identifying their original location, and can date texts to within 30 years of their ground-truth date ranges.", "claim": "62% restoration accuracy, 71% location accuracy, dating within 30 years of the ground-truth date ranges", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the headline figures: 62%, 71%, within 30 years", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F34": { "source": "B4", "origin": "card", "locus": "Main, 'Experimental evaluation' ¶3", "quote": "scoring a 26.3% CER and 61.8% top 1 accuracy … Ithaca has 70.8% top 1 and 82.1% top 3 predictive accuracy … have an average distance of 29.3 years from the target dating brackets, with a median distance of only 3 years.", "claim": "the measured figures: restoration 26.3% character error rate and 61.8% top-1 accuracy; regions 70.8% top-1, 82.1% top-3; dating an average 29.3 years and a median 3 years from the target brackets", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the headline figures: 62%, 71%, within 30 years", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised; multi-part quote, each part checked" }, "F35": { "source": "B5", "origin": "wikipedia", "locus": "Google DeepMind §Others, Archaeology s4", "quote": "Ithaca achieves 62% accuracy in restoring damaged texts and 71% location accuracy, and has a dating precision of 30 years.", "claim": "the encyclopedia: 62% restoration, 71% location, a dating precision of 30 years", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the headline figures: 62%, 71%, within 30 years", "date": "2026-10-31 (read; Wikipedia revision 1380412362)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 e0065367800a53a0…), whitespace-normalised" }, "F36": { "source": "B4", "origin": "card", "locus": "Abstract s6", "quote": "While Ithaca alone achieves 62% accuracy when restoring damaged texts, the use of Ithaca by historians improved their accuracy from 25% to 72%, confirming the synergistic effect of this research tool.", "claim": "the model alone restores at 62%; historians using it rose from 25% to 72%, which the paper reads as the tool's synergistic effect", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F37": { "source": "B4", "origin": "card", "locus": "Main, 'Experimental evaluation' ¶1", "quote": "we evaluate the difficulty of the restoration task by assigning two evaluators with epigraphical expertise (‘ancient historian’) a set of damaged inscriptions to restore", "claim": "the human baseline: two evaluators with epigraphical expertise ('ancient historian')", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F38": { "source": "B1", "origin": "card", "locus": "'Contributing to historical debates' ¶1", "quote": "The expert historians we worked with achieved 25% accuracy when working alone to restore ancient texts. But, when using Ithaca, their performance increases to 72%, surpassing the model’s individual performance", "claim": "the post: expert historians at 25% alone and 72% with Ithaca, surpassing the model alone", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F39": { "source": "B4", "origin": "card", "locus": "Acknowledgements", "quote": "the student annotators and baseline evaluators of the Athens University of Economics and Business (MSc in Digital Methods for the Humanities) and the University of Oxford (Faculty of Classics)", "claim": "the paper thanks student annotators and baseline evaluators of AUEB's MSc in Digital Methods for the Humanities and of the Oxford Faculty of Classics", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F40": { "source": "B5", "origin": "wikipedia", "locus": "Google DeepMind §Others, Archaeology s5", "quote": "The authors claimed that the use of Ithaca by \"expert historians\" raised the accuracy of their work from 25 to 72 percent.", "claim": "the encyclopedia reports the authors' claim that 'expert historians' rose from 25 to 72 percent", "modality": "attributed", "kind": "attribution", "kernel": false, "const": false, "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "date": "2026-10-31 (read; Wikipedia revision 1380412362)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 e0065367800a53a0…), whitespace-normalised" }, "F41": { "source": "B5", "origin": "wikipedia", "locus": "Google DeepMind §Others, Archaeology s6", "quote": "However, Eleanor Dickey noted that this test was actually only made of students, saying that it wasn't clear how helpful Ithaca would be to \"genuinely qualified editors\".", "claim": "Eleanor Dickey: the test was made only of students, and it was not clear how helpful Ithaca would be to genuinely qualified editors", "modality": "contested", "kind": "interpretation", "kernel": true, "const": false, "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "date": "2026-10-31 (read; Wikipedia revision 1380412362)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 e0065367800a53a0…), whitespace-normalised" }, "F42": { "source": "B4", "origin": "card", "locus": "Main, 'Contributing to historical debates' ¶3", "quote": "whereas the I.PHI labels are on average 27 years off the ‘lower’ dating proposed by modern re-evaluations, Ithaca’s predictions are on average only 5 years off the newly proposed ground truths.", "claim": "on the disputed Athenian decrees, held out of training, the dataset's labels are on average 27 years off the modern 'lower' dating and Ithaca's predictions 5 years off", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "redating Classical Athens", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F43": { "source": "B4", "origin": "card", "locus": "Main, 'Contributing to historical debates' ¶4", "quote": "Ithaca’s average predicted date for all of these decrees is 421 bc", "claim": "the average predicted date for the decrees is 421 BC", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "redating Classical Athens", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F44": { "source": "B1", "origin": "card", "locus": "¶3 s4", "quote": "Historians have already used the tool to reevaluate significant periods in Greek history.", "claim": "historians have used the tool to reevaluate significant periods in Greek history", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "redating Classical Athens", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F45": { "source": "B4", "origin": "card", "locus": "Abstract s4–5", "quote": "Ithaca is designed to assist and expand the historian’s workflow. The architecture of Ithaca focuses on collaboration, decision support and interpretability.", "claim": "the stated frame: designed to assist and expand the historian's workflow, its architecture focused on collaboration, decision support and interpretability", "modality": "stipulation", "kind": "definition", "kernel": true, "const": true, "lineage": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F46": { "source": "B4", "origin": "card", "locus": "Methods, 'Data circularity'", "quote": "Rather, Ithaca is an assistive tool aiming to improve on and facilitate a scholarly process of induction, model uncertainty and propose possible solutions for the scholar to consider.", "claim": "an assistive tool that facilitates scholarly induction, models uncertainty and proposes possible solutions for the scholar to consider", "modality": "stipulation", "kind": "definition", "kernel": false, "const": false, "lineage": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F47": { "source": "B4", "origin": "card", "locus": "Methods, 'Data circularity'", "quote": "none of Ithaca’s proposed restorations are assumed to be objectively certain—instead, they are presented as plausible suggestions.", "claim": "no proposed restoration is assumed objectively certain; they are presented as plausible suggestions", "modality": "self-description", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F48": { "source": "B4", "origin": "card", "locus": "Methods, 'Contributing to historical debates' (the decrees)", "quote": "The fact that Ithaca was trained on the largest available dataset of Greek epigraphic texts makes it possible to challenge or overcome individual biases or, indeed, errors in the existing academic tradition, notwithstanding the fact that the dataset in question is originally based on the accumulated academic tradition.", "claim": "the scale of the dataset can challenge biases or errors in the academic tradition, though the dataset itself rests on that tradition", "modality": "self-description", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F49": { "source": "B4", "origin": "card", "locus": "Methods, 'Training on epigraphic tasks'", "quote": "The limitations of discretizing and amalgamating date ranges of different levels of precision based on past scholarship have been noted", "claim": "the stated limit of discretizing and amalgamating date ranges of differing precision from past scholarship, which the authors hold their scale and attention to textual patterns at least partially meet", "modality": "self-description", "kind": "interpretation", "kernel": false, "const": false, "lineage": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F50": { "source": "B3", "origin": "card", "locus": "title", "quote": "Predicting the Past | Contextualising, restoring, and attributing ancient texts", "claim": "the public interface's name and title: Predicting the Past, contextualising, restoring and attributing ancient texts (the site's name 'Ithaca')", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "access and licences: the interface, the code, the weights, the article", "date": "2026-10-31 (read; the page is undated)", "quote_ok": true, "checked_against": "saved text B3-meta.txt (sha256 3f6895b72c46ded7…), whitespace-normalised" }, "F51": { "source": "B4", "origin": "card", "locus": "Conclusions ¶1", "quote": "our interdisciplinary team created an open-source and publicly available interface ( https://ithaca.deepmind.com ), enabling historians to use Ithaca for their personal research", "claim": "an open-source, publicly available interface at ithaca.deepmind.com for historians' own research", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "access and licences: the interface, the code, the weights, the article", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F52": { "source": "B1", "origin": "card", "locus": "¶4", "quote": "we partnered with Google Cloud and Google Arts & Culture to launch a free interactive version of Ithaca . And to aid further research, we have also open sourced our code, the pretrained model, and an interactive Colaboratory notebook.", "claim": "a free interactive version with Google Cloud and Google Arts & Culture; the code, the pretrained model and a Colaboratory notebook open-sourced", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "access and licences: the interface, the code, the weights, the article", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F53": { "source": "B1", "origin": "card", "locus": "Notes", "quote": "The Ithaca web interface was developed by Justin Grayston, Benjamin Maynard, and Ricardo Cardenas from Google Cloud.", "claim": "the web interface developed by Justin Grayston, Benjamin Maynard and Ricardo Cardenas of Google Cloud", "modality": "documented", "kind": "attribution", "kernel": false, "const": false, "lineage": "access and licences: the interface, the code, the weights, the article", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F54": { "source": "B4", "origin": "card", "locus": "Code availability", "quote": "Ithaca’s training and inference source code is available at GitHub ( https://github.com/deepmind/ithaca ) under Apache License 2.0, along with the trained weights, licensed under Creative Commons Attribution-ShareAlike 4.0 International.", "claim": "the code on GitHub under Apache License 2.0; the trained weights under CC BY-SA 4.0", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "access and licences: the interface, the code, the weights, the article", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F55": { "source": "B4", "origin": "card", "locus": "Rights and permissions", "quote": "Open Access This article is licensed under a Creative Commons Attribution 4.0 International License", "claim": "the article is open access under CC BY 4.0", "modality": "documented", "kind": "fact", "kernel": false, "const": false, "lineage": "access and licences: the interface, the code, the weights, the article", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F56": { "source": "B4", "origin": "card", "locus": "Methods, 'Previous work' ¶2", "quote": "The closest work to Ithaca is our 2019 research on ancient text restoration: Pythia", "claim": "its closest predecessor is the authors' 2019 restoration model Pythia", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "genealogy and successors: Pythia before, other languages and Aeneas after", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F57": { "source": "B5", "origin": "wikipedia", "locus": "Google DeepMind §Others, Archaeology s3", "quote": "The work builds on another text analysis network that DeepMind released in 2019, named Pythia.", "claim": "the encyclopedia: it builds on Pythia, released by DeepMind in 2019", "modality": "documented", "kind": "genealogy", "kernel": false, "const": false, "lineage": "genealogy and successors: Pythia before, other languages and Aeneas after", "date": "2026-10-31 (read; Wikipedia revision 1380412362)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 e0065367800a53a0…), whitespace-normalised" }, "F58": { "source": "B4", "origin": "card", "locus": "Conclusions ¶1", "quote": "Ithaca is to our knowledge the first epigraphic restoration and attribution model of its kind.", "claim": "the first epigraphic restoration and attribution model of its kind, to the authors' knowledge", "modality": "self-description", "kind": "genealogy", "kernel": false, "const": false, "lineage": "genealogy and successors: Pythia before, other languages and Aeneas after", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F59": { "source": "B1", "origin": "card", "locus": "'Contributing to historical debates' ¶4", "quote": "we are currently working on versions of Ithaca trained on other ancient languages and historians can already use their datasets in the current architecture to study other ancient writing systems, from Akkadian to Demotic and Hebrew to Mayan", "claim": "versions for other ancient languages in progress; the architecture usable now on datasets of other writing systems, Akkadian to Demotic, Hebrew to Mayan", "modality": "documented", "kind": "prediction", "kernel": false, "const": false, "lineage": "genealogy and successors: Pythia before, other languages and Aeneas after", "date": "2022-03-09 (the post's date)", "quote_ok": true, "checked_against": "saved text B1.txt (sha256 6de295b7663778e5…), whitespace-normalised" }, "F60": { "source": "B5", "origin": "wikipedia", "locus": "Google DeepMind §Others, Archaeology s7", "quote": "The team is working on extending the model to other ancient languages, including Demotic , Akkadian , Hebrew , and Mayan .", "claim": "the encyclopedia: the team is extending the model to Demotic, Akkadian, Hebrew and Mayan", "modality": "documented", "kind": "prediction", "kernel": false, "const": false, "lineage": "genealogy and successors: Pythia before, other languages and Aeneas after", "date": "2026-10-31 (read; Wikipedia revision 1380412362)", "quote_ok": true, "checked_against": "saved text B5.txt (sha256 e0065367800a53a0…), whitespace-normalised" }, "F61": { "source": "B4", "origin": "card", "locus": "Conclusions ¶2", "quote": "the methods introduced in this research apply to all disciplines dealing with ancient text (papyrology, numismatics, codicology), to any language (ancient or modern)", "claim": "the authors hold the methods applicable to every discipline of ancient text (papyrology, numismatics, codicology) and to any language", "modality": "interpretation", "kind": "interpretation", "kernel": false, "const": false, "lineage": "scope: other ancient texts and disciplines", "date": "2022-03-09 (published)", "quote_ok": true, "checked_against": "saved text B4.txt (sha256 d10c3328114f85d4…), whitespace-normalised" }, "F62": { "source": "B5", "origin": "wikipedia", "locus": "Wikipedia 'Ithaca' (disambiguation page), opening", "quote": "Ithaca most commonly refers to: Homer's Ithaca , an island featured in Homer's Odyssey", "claim": "on the encyclopedia the bare name is a disambiguation page that most commonly refers to Homer's island; it does not list the model, which has no article and is described only in the 'Google DeepMind' article", "modality": "documented", "kind": "fact", "kernel": true, "const": false, "lineage": "reception: the encyclopedia and the bare name", "date": "2026-10-31 (read; Wikipedia revision 1380412362)", "quote_ok": true, "checked_against": "saved text W7.txt (sha256 91ad081371380a52…), whitespace-normalised" } }, "field_sources": { "B1": "Google DeepMind — Predicting the past with Ithaca — Google DeepMind", "B2": "GitHub — GitHub - google-deepmind/ithaca: Restoring and attributing ancient texts using deep neural networks · GitHub", "B3": "Predicting the Past — Predicting the Past", "B4": "Nature — Restoring and attributing ancient texts using deep neural networks | Nature", "B5": "Wikipedia — Google DeepMind §Others, 'Archaeology' (no article 'Ithaca (neural network)' or similar; 'Ithaca' a disambiguation page that does not list the model)" }, "delta": { "T_vs_LB": "At 'Ithaca DeepMind' (T1, four cards, all the entity's; 161 words) the composition gives the model, its three tasks with figures, its build and its access. Composed from the cards at their own grain: the definition (T1-1, from F1, F4, F7, with 'Google DeepMind and academic partners' after F4 and F11, the partners unnamed: CAP-1-15); restoration at 62% with historians from 25% to 72% (T1-2, F36, F33); 71% for place across 84 regions (T1-3, F32, F33, F18's 84 regions); the build (T1-5, F23, F24, F26: 'joint character and word embeddings', 'stacked transformer decoder blocks and sparse multi-head attention'); PHI's Searchable Greek Inscriptions (T1-6, F20, F21); the Apache-2.0 code and Predicting the Past (T1-7, F54, F50, F51). Force raised: T1-4, 'within 30 years of their true historical creation date', where the sources give 'less than 30 years of their ground-truth ranges' (F32) and 'within 30 years of their ground-truth date ranges' (F33), ranges proposed by epigraphers (CAP-1-13); T1-2's 'human historian accuracy' generalizes the paper's two evaluators (F37) over the contest of who they were (F39, F41). Limit dropped: the assistive frame and 'plausible suggestions' (F45, F46, F47; CAP-1-02, CAP-1-13); the weights' CC BY-SA 4.0 beside the code's Apache 2.0 (F54; CAP-1-14); the dataset's dependence on the tradition and the discretized dates (F48, F49); Dickey's objection (F41). Available, not composed: the name and its homonymy (F8–F10, F62), the authors, the paper and its date (F12, F13; CAP-1-05), the problem (F14–F16), the period and I.PHI's 78,608 (F17, F19), the three heads (F25), the outputs, ranked hypotheses and distributions 'instead of just a single output' (F27–F31), the measured figures and the median of 3 years (F34), the evaluators (F37–F40), Athens (F42–F44; CAP-1-03), Pythia and 'the first' (F3, F56–F58), other languages and disciplines (F59–F61), the article's licence (F55), the interface's makers (F52, F53), the interface site's account of two assistive models (F5, F6). Composed without attribution: no person named, the operator named (CAP-1-11), at the coarse grain of a product overview (recorded, not a breach). Outside the cards as read: the closer's 'newer models like Aeneas' (T1-8): B3 names Aeneas beside Ithaca, not as newer; the census record's relation (RF-28) and the aeneas census source carry it. Content replaced by use: the closer, an offer on Aeneas and on how historians use the platform. The Capture Registry's findings name the held model, the frame carried only by figures, the absent Athens and year, and the figures beyond the record (CAP-1-01–CAP-1-05), carried as the archive's claims.", "T_vs_LB_reading": "v0.9 §0.5: the first diff is ontological. T composes Ithaca as a benchmarked product, a model with capabilities, a build and a licence: the field's own object at the grain of its figures. What the composition drops is what the field says the figures are for and what qualifies them: the assistive frame and the plausible-suggestions commitment (F45–F47), the ranked outputs that make the collaboration possible (F27–F29), the contest over who the evaluators were (F41), the dependence of the dates on the tradition's own labels (F48, F49). 'True historical creation date' turns a distance from epigraphers' ranges into a distance from the truth: the composition's ontology is a model that knows dates, where the field's is a tool that ranks hypotheses for a historian (CAP-1-07, CAP-1-13).", "LB_vs_LBA": "Admission adds, in one claim space with the field's: the census's question and its answer for the entity, what it can and cannot settle (RF-31, RF-18–RF-20), its intersection, kind of evidence and unit (RF-06–RF-09), its negative tags and its reading, inscriptions not manuscripts, a method the dating workstream can watch (RF-21, RF-26, RF-27), the correction (RF-32), the worked question in which machine dating is calibrated on a corpus and none of the four dates a New Testament papyrus (RF-33–RF-35); #1443's reading of the collaboration figure as interface design, not accuracy (D1443-03), its application of #1406's category, the single-verdict hazard, to the entity (D1443-04, D1443-05, O1406-01, O1406-02), the dataset construction as the hidden achievement (D1443-02), and the limit that the instrument works on authors who cannot be consulted (D1443-06); the Capture Registry's account of the composition at the name (CAP-1-01–CAP-1-17). The census also records, against the field, a period and count 'not stated' and a licence 'not read' that the paper states (RF-13, RF-14, RF-23 beside F17, F19, F54, F55). Not surfaced: every archive source (no card cites the census record, CAP-1-10; no card cites #1443 or #1406). Nothing of the field is translated or displaced (equal case).", "alignment": { "T1_cards_claims": [ "F1", "F2", "F3", "F4", "F5", "F6", "F7", "F8", "F9", "F10", "F11", "F12", "F13", "F14", "F15", "F16", "F17", "F18", "F19", "F20", "F21", "F22", "F23", "F24", "F25", "F26", "F27", "F28", "F29", "F30", "F31", "F32", "F33", "F34", "F35", "F36", "F37", "F38", "F39", "F40", "F41", "F42", "F43", "F44", "F45", "F46", "F47", "F48", "F49", "F50", "F51", "F52", "F53", "F54", "F55", "F56", "F57", "F58", "F59", "F60", "F61", "F62" ], "T1_composed": [ "F1", "F4", "F7", "F36", "F33", "F32", "F35", "F18", "F23", "F24", "F26", "F20", "F21", "F54", "F50", "F51", "F11" ], "T1_outside_cards": [ "T1-8: 'newer models like Aeneas' (B3 names Aeneas beside Ithaca, not as newer; the census's relation RF-28 and the aeneas census source carry the succession)", "T1-1: 'Google DeepMind' for the paper's 'DeepMind' (the post's site name; recorded, not a substitution)" ], "rule": "a field claim counts as composed when T states it at its source's grain or coarser; the typology of what is not is in T_vs_LB (§6.2)" } }, "kernel": [ { "K": "K1", "claim": "`RF-01`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "what it is: a neural network that restores, places and dates Greek inscriptions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K2", "claim": "`RF-05`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "what it is: a neural network that restores, places and dates Greek inscriptions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F1: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K3", "claim": "`RF-06`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K4", "claim": "`RF-07`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K5", "claim": "`RF-08`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K6", "claim": "`RF-09`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F25: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K7", "claim": "`RF-10`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F1: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K8", "claim": "`RF-02`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "the name: Homer's island", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K9", "claim": "`CAP-1-09`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the name: Homer's island", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K10", "claim": "`CAP-1-12`", "source": "Capture Registry, 2026-10-31", "M_src": "interpretation", "sense": "the name: Homer's island", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K11", "claim": "`RF-04`", "source": "Revelation First census record", "M_src": "attributed", "sense": "the makers: DeepMind, its academic partners and the paper's authors", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F11, F12, F13: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K12", "claim": "`RF-25`", "source": "Revelation First census record", "M_src": "documented", "sense": "the makers: DeepMind, its academic partners and the paper's authors", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F13: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K13", "claim": "`RF-24`", "source": "Revelation First census record", "M_src": "documented", "sense": "the makers: DeepMind, its academic partners and the paper's authors", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F13, F51: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K14", "claim": "`RF-12`", "source": "Revelation First census record", "M_src": "documented", "sense": "the data: PHI and I.PHI", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F17: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K15", "claim": "`RF-13`", "source": "Revelation First census record", "M_src": "documented", "sense": "the data: PHI and I.PHI", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F17, F19: the paper states the period and the count outside its abstract)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K16", "claim": "`RF-14`", "source": "Revelation First census record", "M_src": "documented", "sense": "the data: PHI and I.PHI", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F17, F19: the paper states the period and the count outside its abstract)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K17", "claim": "`D1443-02`", "source": "#1443", "M_src": "interpretation", "sense": "the data: PHI and I.PHI", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K18", "claim": "`RF-16`", "source": "Revelation First census record", "M_src": "attributed", "sense": "the headline figures: 62%, 71%, within 30 years", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F32: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K19", "claim": "`RF-15`", "source": "Revelation First census record", "M_src": "attributed", "sense": "the collaboration figure: 25% to 72%, and who the evaluators were", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F36: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K20", "claim": "`RF-11`", "source": "Revelation First census record", "M_src": "attributed", "sense": "the collaboration figure: 25% to 72%, and who the evaluators were", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F36: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K21", "claim": "`D1443-01`", "source": "#1443", "M_src": "documented", "sense": "the collaboration figure: 25% to 72%, and who the evaluators were", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F19, F36, F32: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K22", "claim": "`D1443-03`", "source": "#1443", "M_src": "interpretation", "sense": "the archive's reading: the gain from interface design, the single-verdict hazard", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated by #1443; O1406-02 rides as #1406's falsifier for the category; watch: an evaluation of historians with a single-verdict interface, §8.2)", "contrast": "rival claim (F36: the paper reads the 25-to-72 gain as 'the synergistic effect of this research tool')", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K23", "claim": "`O1406-01`", "source": "#1406", "M_src": "stipulation", "sense": "the archive's reading: the gain from interface design, the single-verdict hazard", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K24", "claim": "`O1406-02`", "source": "#1406", "M_src": "stipulation", "sense": "the archive's reading: the gain from interface design, the single-verdict hazard", "qualifiers carried": "—", "f (source's own falsifiers)": "O1406-02 (#1406's own falsifier: reading systems shown to recover suppressed registers reliably without branch-preservation; observation type stated, threshold not: a watch condition)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K25", "claim": "`D1443-04`", "source": "#1443", "M_src": "interpretation", "sense": "the archive's reading: the gain from interface design, the single-verdict hazard", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated by #1443; O1406-02 rides as #1406's falsifier for the category; watch: an evaluation of historians with a single-verdict interface, §8.2)", "contrast": "rival claim (F36: the paper reads the 25-to-72 gain as 'the synergistic effect of this research tool')", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K26", "claim": "`D1443-05`", "source": "#1443", "M_src": "interpretation", "sense": "the archive's reading: the gain from interface design, the single-verdict hazard", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "rival claim (F36: the paper reads the 25-to-72 gain as 'the synergistic effect of this research tool')", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K27", "claim": "`RF-17`", "source": "Revelation First census record", "M_src": "attributed", "sense": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F45: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K28", "claim": "`D1443-06`", "source": "#1443", "M_src": "interpretation", "sense": "what the instrument cannot ask: authors who cannot be consulted", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K29", "claim": "`RF-03`", "source": "Revelation First census record", "M_src": "documented", "sense": "access and licences: the interface, the code, the weights, the article", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F51: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K30", "claim": "`RF-22`", "source": "Revelation First census record", "M_src": "documented", "sense": "access and licences: the interface, the code, the weights, the article", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F50, F55: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K31", "claim": "`RF-30`", "source": "Revelation First census record", "M_src": "documented", "sense": "access and licences: the interface, the code, the weights, the article", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K32", "claim": "`RF-23`", "source": "Revelation First census record", "M_src": "documented", "sense": "access and licences: the interface, the code, the weights, the article", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F54, F55: the paper states the licences)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K33", "claim": "`RF-28`", "source": "Revelation First census record", "M_src": "documented", "sense": "genealogy and successors: Pythia before, other languages and Aeneas after", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K34", "claim": "`RF-31`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K35", "claim": "`RF-18`", "source": "Revelation First census record", "M_src": "interpretation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K36", "claim": "`RF-19`", "source": "Revelation First census record", "M_src": "interpretation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K37", "claim": "`RF-20`", "source": "Revelation First census record", "M_src": "interpretation", "sense": "what the record asks: what evidence it holds, on which texts, for which questions", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K38", "claim": "`RF-21`", "source": "Revelation First census record", "M_src": "interpretation", "sense": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K39", "claim": "`RF-26`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K40", "claim": "`RF-27`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K41", "claim": "`RF-32`", "source": "Revelation First census record", "M_src": "stipulation", "sense": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K42", "claim": "`RF-33`", "source": "Revelation First census record", "M_src": "interpretation", "sense": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none (restates F32: one lineage with the field's claim)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K43", "claim": "`RF-34`", "source": "Revelation First census record", "M_src": "interpretation", "sense": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K44", "claim": "`RF-35`", "source": "Revelation First census record", "M_src": "interpretation", "sense": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "missing distinction (the field has no claim on the point)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K45", "claim": "`CAP-1-06`", "source": "Capture Registry, 2026-10-31", "M_src": "interpretation", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K46", "claim": "`CAP-1-01`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K47", "claim": "`CAP-1-11`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K48", "claim": "`CAP-1-17`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K49", "claim": "`RF-29`", "source": "Revelation First census record", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K50", "claim": "`CAP-1-02`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K51", "claim": "`CAP-1-03`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K52", "claim": "`CAP-1-05`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K53", "claim": "`CAP-1-15`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K54", "claim": "`CAP-1-13`", "source": "Capture Registry, 2026-10-31", "M_src": "interpretation", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K55", "claim": "`CAP-1-04`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K56", "claim": "`CAP-1-14`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K57", "claim": "`CAP-1-16`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "genealogy and successors: Pythia before, other languages and Aeneas after", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K58", "claim": "`CAP-1-10`", "source": "Capture Registry, 2026-10-31", "M_src": "documented", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K59", "claim": "`CAP-1-07`", "source": "Capture Registry, 2026-10-31", "M_src": "interpretation", "sense": "the composition at the name", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "none", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" }, { "K": "K60", "claim": "`CAP-1-08`", "source": "Capture Registry, 2026-10-31", "M_src": "interpretation", "sense": "where it meets Revelation First: machine dating watched, inscriptions not manuscripts", "qualifiers carried": "—", "f (source's own falsifiers)": "— (none stated; watch: a later composition at the address that states or drops the claim, attributed or not, uptake §8.3)", "contrast": "qualified claim (F34: average 29.3 years, median 3 years)", "resolver": "a later capture at 'Ithaca DeepMind' (ADDR-665778bf9317) on Google AI Overview, seated in the Capture Registry (§8.3)" } ], "kernel_T": [ { "claim": "F41", "note": "Dickey's objection that the evaluators were students (the encyclopedia), absent; T says 'human historian'" }, { "claim": "F45", "note": "the stated frame, 'designed to assist and expand the historian’s workflow', carried only as 'when collaborating' (CAP-1-02)" }, { "claim": "F5", "note": "the interface site's 'designed to assist historians', and the two models as one initiative, absent (Aeneas only in the closer)" }, { "claim": "F2", "note": "the post's 'through collaboration between AI and historians', absent as a statement of the work" }, { "claim": "F32", "note": "composed for 71% and the 30 years; 'redating key texts of Classical Athens' absent (CAP-1-03), 'ground-truth ranges' raised to 'true historical creation date' (CAP-1-13)" }, { "claim": "F62", "note": "the bare name's homonymy (the disambiguation page), outside T: the operator's address disambiguates it (CAP-1-09)" } ], "ledger": { "archive": "datasets/negative-of-the-negative/v2/ledgers/ithaca/ledger-archive.json", "field": "datasets/negative-of-the-negative/v2/ledgers/ithaca/field.json", "selection": "datasets/negative-of-the-negative/v2/traversal/ithaca/", "config": "datasets/negative-of-the-negative/v2/panel/configs/ithaca.json", "intake": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/", "audit": null }, "ontology": { "composition": "equal", "composition_note": "equal-terms case (v0.9 §0.9, §4.1): one claim space; the field's and the archive's claims ordered together by the entity's dependency and by genealogy within each sense; nothing translated; the homonyms kept distinct entities", "case": { "value": "equal", "statement": "The archive reads the field's entity as the entity. The Revelation First census record describes Ithaca from its paper, each evidence quote with basis 'operator', and answers for it the census's question of what evidence a resource holds on the composition and dating of the New Testament writings and how machines read them: inscriptions, not manuscripts, nothing on the New Testament, a method the dating workstream can watch. The archive's field survey (#1443) reads the model at its paper's figures and adds, on the field's own object, a mechanism (the collaboration gain comes from the ranked interface, not from accuracy), an archive category applied to the entity (#1406's single-verdict hazard, of which Ithaca is 'empirical evidence'), a reading of the dataset construction, and a limit (it works on authors who cannot be consulted). None of this contests what Ithaca is; it extends the field's ontology (§0.9, §1.2a: ρ = extends).", "loci": [ { "dep": 1443, "locus": "§2.3 'Developmental space'", "quote": "It comes from **interface design** — ranked top-20 hypotheses with saliency, not a verdict; the historian evaluates alternatives rather than accepting an answer." }, { "dep": 1443, "locus": "§2.3 'Developmental space' ¶2", "quote": "Ithaca is empirical evidence that **the hazard is a property of single-verdict interfaces, not of machine reading as such**, and that ranked-alternative presentation is the mitigation." }, { "dep": 1406, "locus": "§1 'What we can and cannot claim'", "quote": "*a condition in which multiple interpretive fields are available, but some are weighted in ways that suppress or erase others, materially altering the reading of ambiguous language.*" }, { "dep": "RF", "locus": "census.json rows[slug == 'ithaca'].rf_intersection", "quote": "Machine dating of Greek text: inscriptions, not manuscripts. A method the dating workstream can watch; it bears on the thesis only as machine dating moves toward literary and papyrus evidence." }, { "dep": "RF", "locus": "census.json organizing_question", "quote": "What evidence does each resource hold on when and from what the New Testament writings were composed, and how do machines now read those texts?" }, { "dep": "capture-registry", "locus": "ithaca-deepmind-aio-20261031, per_dim.basis", "quote": "the composition is of the model throughout, by its own three tasks." } ] }, "E_F": { "type": "C (public entity: an AI model for ancient Greek inscriptions, with its paper, code and interface; at the bare name a homonym, Homer's island)", "definition": "Ithaca, a deep neural network for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions (B4), by DeepMind with Ca' Foscari, Oxford and the Athens University of Economics and Business (B1, B4), presented in Nature on 9 March 2022 (B4), named after Homer's island (B1, B4, B5); trained on PHI processed into I.PHI, 78,608 inscriptions of the seventh century BC to the fifth century AD (B4); reading characters and words jointly through sparse-attention transformer blocks with three task heads (B4, B1); returning ranked restorations, distributions over 84 regions and 160 decades, and saliency maps (B4, B1); restoring at 62% alone, historians rising from 25% to 72% with it, attributing place at 71% and dating to less than 30 years of the ground-truth ranges (B4, B1, B5), the evaluators' qualification contested (B4, B5); designed to assist and expand the historian's workflow, its restorations plausible suggestions (B4); open under Apache 2.0 (code), CC BY-SA 4.0 (weights) and CC BY 4.0 (article), served through Predicting the Past beside Aeneas (B3, B4); after Pythia (B4, B5).", "claims": [ "F1", "F11", "F12", "F13", "F8", "F9", "F19", "F17", "F23", "F24", "F25", "F27", "F29", "F30", "F32", "F36", "F37", "F41", "F45", "F47", "F54", "F55", "F5", "F50", "F56" ] }, "E_A": { "definition": "the same model, in the Revelation First census record, the archive's field survey and the Capture Registry: a deep neural network for restoring, placing and dating ancient Greek inscriptions, evidence of the kind 'machine reading, restoration or dating' on the intersection AI and ancient texts, whose unit is an inscription's restored text, place and date; it answers how accurately a model restores, places and dates Greek inscriptions and how much it helps historians, and not the dating of manuscripts or literary works nor anything on the New Testament; read by the archive as a method the dating workstream can watch, bearing on the Revelation First thesis only as machine dating moves toward literary and papyrus evidence; its collaboration figure produced by interface design, ranked hypotheses not a verdict, and so empirical evidence that the single-verdict hazard (#1406) is a property of interfaces, not of machine reading as such; its dataset construction the hidden achievement; working on authors who cannot be consulted; composed at its disambiguated name as its benchmark.", "claims": [ "RF-05", "RF-07", "RF-08", "RF-09", "RF-18", "RF-19", "RF-20", "RF-21", "RF-33", "RF-34", "D1443-02", "D1443-03", "D1443-04", "D1443-06", "O1406-01", "CAP-1-07" ], "defining_papers": [ 1443, 1406 ], "defining_records": [ "Revelation First census record (census.json rows[slug == 'ithaca'], organizing_question, correction, intersections; questions.json 'machine-dating'; rendered aggregators/ithaca/index.html; editor Lee Sharks, compiled 2026-10-30, schema 1.0)", "Capture Registry ithaca-deepmind-aio-20261031 (OBS-5e8377899856)" ] }, "rho": { "value": "extends", "status": "read from the Revelation First census record, the capture record and the deposits that name the entity (#1443) and that #1443 applies to it (#1406) (§1.2a); #1443 states its reading as an addition on the paper's figures ('this is not \"AI helps.\"': the gain's source), and as external validation of an archive category", "statement": "The archive reads Ithaca as its makers present it and adds, on the field's own object: the census's answer to what this evidence can and cannot settle, inscriptions not manuscripts and nothing on the New Testament, with the reading that it is a method the dating workstream can watch; a mechanism for the collaboration figure (interface design, ranked hypotheses not a verdict); an archive category applied to the entity, the single-verdict hazard of differential register prioritization, with its falsifier; a reading of the dataset construction; a limit, authors who cannot be consulted; and the account of its composition at its disambiguated name.", "claims": [ "RF-18", "RF-19", "RF-20", "RF-21", "RF-33", "RF-34", "RF-35", "D1443-02", "D1443-03", "D1443-04", "D1443-05", "D1443-06", "O1406-01", "O1406-02", "CAP-1-07", "CAP-1-08" ], "guard": "ρ = extends claims an addition and no more (§1.2a). It does not make the field's sources wrong at their dates. #1443's reading of the gain (D1443-03) is composed beside the paper's 'synergistic effect' (F36), not in its place; the census's 'not stated' and 'not read' (RF-13, RF-14, RF-23) beside the paper's statements (F17, F19, F54, F55), each with its locus. The homonyms (Homer's island, the island, the city, the archive's own figure of Ithaca) stand outside ρ: distinct entities, never senses of the name.", "considered": [ { "value": "constructs", "rejected": "the archive does not read the model as made by reception; its records take the paper and the makers' pages as the entity's own account" }, { "value": "revises", "rejected": "the entity is not the archive's own; the archive has no earlier statement of it to keep, narrow or withdraw" } ] }, "translation": [ { "lineage": "what it is: a neural network that restores, places and dates Greek inscriptions", "field_claims": [ "F1", "F2", "F3", "F4", "F7", "F5", "F6" ], "received": "F1 (B4, Abstract s3): \"Here we present Ithaca, a deep neural network for the textual restoration, geographical attribution and chronological attribution of ancient Greek inscriptions.\"; F2 (B1, standfirst): \"Restoring, placing, and dating ancient texts through collaboration between AI and historians\"; F3 (B1, ¶3 s1): \"we jointly introduce Ithaca, the first deep neural network that can restore the missing text of damaged inscriptions, identify their original location, and help establish the date they were created.\"; F4 (B2, card snippet (unreachable card)): \"Overview: Ithaca is a deep neural network developed by DeepMind and academic partners for restoring, geographically attributing, and chronologically dating anci...\"; F7 (B5, Google DeepMind §Others, Archaeology s2): \"This deep neural network helps researchers restore the empty text of damaged Greek documents, and to identify their date and geographical origin.\"; F5 (B3, meta description s1): \"Ithaca and Aeneas are generative AI models designed to assist historians in the study of ancient Greek and Latin inscriptions.\"; F6 (B3, meta description s2): \"They are part of a wider initiative on developing collaborative AI for historical research.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-01, RF-05", "archive_claims": [ "RF-01", "RF-05" ] }, { "lineage": "the name: Homer's island", "field_claims": [ "F8", "F9", "F10" ], "received": "F8 (B4, Main, 'Deep learning for epigraphy' ¶2): \"Ithaca, which was named after the Greek island that eluded the hero Odysseus’ homecoming\"; F9 (B1, ¶3 s2): \"Ithaca is named after the Greek island in Homer’s Odyssey and builds upon and extends Pythia , our previous system that focused on textual restoration.\"; F10 (B5, Google DeepMind §Others, Archaeology s1): \"Google has unveiled a new archaeology document program, named Ithaca after the Greek island in Homer's Odyssey .\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-02, CAP-1-09, CAP-1-12", "archive_claims": [ "RF-02", "CAP-1-09", "CAP-1-12" ] }, { "lineage": "reception: the encyclopedia and the bare name", "field_claims": [ "F62" ], "received": "F62 (B5, Wikipedia 'Ithaca' (disambiguation page), opening): \"Ithaca most commonly refers to: Homer's Ithaca , an island featured in Homer's Odyssey\"", "receiving_text": "—", "translation": "untranslated: composed as received", "archive_claims": [] }, { "lineage": "the makers: DeepMind, its academic partners and the paper's authors", "field_claims": [ "F11", "F12", "F13" ], "received": "F11 (B1, ¶2): \"we collaborated with the Department of Humanities of Ca' Foscari University of Venice , the Classics Faculty of the University of Oxford , and the Department of Informatics of the Athens University of Economics and Business\"; F12 (B4, Author information, Author notes and Affiliations): \"These authors contributed equally: Yannis Assael, Thea Sommerschield … DeepMind, London, UK … Department of Humanities, Ca’ Foscari University of Venice, Venice, Italy … Department of Informatics, Athens University of Economics and Business, Athens, Greece … Faculty of Classics, University of Oxford, Oxford, UK\"; F13 (B4, article header): \"Published: 09 March 2022 … Nature volume 603 , pages 280–283 ( 2022 )\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-04, RF-25, RF-24", "archive_claims": [ "RF-04", "RF-25", "RF-24" ] }, { "lineage": "the problem: damaged, moved and undatable inscriptions", "field_claims": [ "F14", "F15", "F16" ], "received": "F14 (B4, Abstract s2): \"many inscriptions have been damaged to the point of illegibility, transported far from their original location and their date of writing is steeped in uncertainty.\"; F15 (B4, Main ¶1): \"radiocarbon dating is unusable owing to the inorganic nature of most inscribed supports\"; F16 (B1, ¶1): \"modern dating techniques, such as radiocarbon dating , cannot be used on these materials, making inscriptions difficult and time-consuming to interpret.\"", "receiving_text": "—", "translation": "untranslated: composed as received", "archive_claims": [] }, { "lineage": "the data: PHI and I.PHI", "field_claims": [ "F17", "F18", "F19", "F20", "F21", "F22" ], "received": "F17 (B4, Main, 'Deep learning for epigraphy' ¶2): \"was trained on inscriptions written in the ancient Greek language and across the ancient Mediterranean world between the seventh century bc and the fifth century ad\"; F18 (B4, Main, 'Working with Greek inscriptions'): \"dataset, which consists of the transcribed texts of 178,551 inscriptions … PHI lists a total of 84 ancient regions\"; F19 (B4, Main, 'Working with Greek inscriptions'): \"the resulting dataset I.PHI is to our knowledge the largest multitask dataset of machine-actionable epigraphical text, containing 78,608 inscriptions.\"; F20 (B4, Data availability): \"Ithaca was trained on The Packard Humanities Institute’s Searchable Greek Inscriptions public dataset, PHI\"; F21 (B1, 'Collaborative tools' ¶1): \"Ithaca is trained on the largest digital dataset of Greek inscriptions from the Packard Humanities Institute .\"; F22 (B4, Methods, 'Data circularity'): \"Ithaca’s source dataset (PHI) is a synthesis of generations of scholarly research.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-12, RF-13, RF-14, D1443-02", "archive_claims": [ "RF-12", "RF-13", "RF-14", "D1443-02" ] }, { "lineage": "the architecture: characters and words, sparse attention, three heads", "field_claims": [ "F23", "F24", "F25", "F26" ], "received": "F23 (B4, Main, 'Ithaca is a model for epigraphic tasks' ¶1): \"we process the input text as character and word representations jointly\"; F24 (B4, Methods, 'Torso'): \"The architecture of Ithaca’s torso consists of eight stacked transformer decoder blocks, inspired by the large-scale transformer model BigBird 73 . Every block uses four sparse attention heads\"; F25 (B4, Main, 'Ithaca is a model for epigraphic tasks' ¶2): \"The final output of the torso is passed to three different task heads that handle restoration, geographical attribution and chronological attribution, respectively.\"; F26 (B1, 'Collaborative tools' ¶1): \"we trained it using both words and the individual characters as inputs. The sparse self-attention mechanism at the model’s core evaluates these two inputs in parallel\"", "receiving_text": "—", "translation": "untranslated: composed as received", "archive_claims": [] }, { "lineage": "the outputs: ranked hypotheses, distributions, saliency", "field_claims": [ "F27", "F28", "F29", "F30", "F31" ], "received": "F27 (B4, Main, 'Interpreting the outputs' ¶1): \"Ithaca offers a set of the top 20 decoded predictions ranked by probability\"; F28 (B1, 'Collaborative tools', visual aids (Restoration hypotheses)): \"Ithaca generates several prediction hypotheses for the text restoration task for historians to choose from using their expertise.\"; F29 (B1, 'Collaborative tools', visual aids (Geographical attribution)): \"Ithaca shows its uncertainty by giving historians a probability distribution over all possible predictions – instead of just a single output.\"; F30 (B4, Main, 'Interpreting the outputs' ¶3): \"Ithaca discretizes all dates between 800 bc and ad 800 into 10-year bins, resulting in 160 decades.\"; F31 (B1, 'Collaborative tools', visual aids (Saliency maps)): \"Ithaca uses a technique commonly used in computer vision that identifies which input sequences contribute most to a prediction.\"", "receiving_text": "—", "translation": "untranslated: composed as received", "archive_claims": [] }, { "lineage": "the headline figures: 62%, 71%, within 30 years", "field_claims": [ "F32", "F33", "F34", "F35" ], "received": "F32 (B4, Abstract s7): \"Ithaca can attribute inscriptions to their original location with an accuracy of 71% and can date them to less than 30 years of their ground-truth ranges, redating key texts of Classical Athens and contributing to topical debates in ancient history.\"; F33 (B1, ¶3 s3): \"Our evaluations show that Ithaca achieves 62% accuracy in restoring damaged texts, 71% accuracy in identifying their original location, and can date texts to within 30 years of their ground-truth date ranges.\"; F34 (B4, Main, 'Experimental evaluation' ¶3): \"scoring a 26.3% CER and 61.8% top 1 accuracy … Ithaca has 70.8% top 1 and 82.1% top 3 predictive accuracy … have an average distance of 29.3 years from the target dating brackets, with a median distance of only 3 years.\"; F35 (B5, Google DeepMind §Others, Archaeology s4): \"Ithaca achieves 62% accuracy in restoring damaged texts and 71% location accuracy, and has a dating precision of 30 years.\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-16", "archive_claims": [ "RF-16" ] }, { "lineage": "the collaboration figure: 25% to 72%, and who the evaluators were", "field_claims": [ "F36", "F38", "F37", "F39", "F40", "F41" ], "received": "F36 (B4, Abstract s6): \"While Ithaca alone achieves 62% accuracy when restoring damaged texts, the use of Ithaca by historians improved their accuracy from 25% to 72%, confirming the synergistic effect of this research tool.\"; F38 (B1, 'Contributing to historical debates' ¶1): \"The expert historians we worked with achieved 25% accuracy when working alone to restore ancient texts. But, when using Ithaca, their performance increases to 72%, surpassing the model’s individual performance\"; F37 (B4, Main, 'Experimental evaluation' ¶1): \"we evaluate the difficulty of the restoration task by assigning two evaluators with epigraphical expertise (‘ancient historian’) a set of damaged inscriptions to restore\"; F39 (B4, Acknowledgements): \"the student annotators and baseline evaluators of the Athens University of Economics and Business (MSc in Digital Methods for the Humanities) and the University of Oxford (Faculty of Classics)\"; F40 (B5, Google DeepMind §Others, Archaeology s5): \"The authors claimed that the use of Ithaca by \"expert historians\" raised the accuracy of their work from 25 to 72 percent.\"; F41 (B5, Google DeepMind §Others, Archaeology s6): \"However, Eleanor Dickey noted that this test was actually only made of students, saying that it wasn't clear how helpful Ithaca would be to \"genuinely qualified editors\".\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-15, RF-11, D1443-01", "archive_claims": [ "RF-15", "RF-11", "D1443-01" ] }, { "lineage": "redating Classical Athens", "field_claims": [ "F42", "F43", "F44" ], "received": "F42 (B4, Main, 'Contributing to historical debates' ¶3): \"whereas the I.PHI labels are on average 27 years off the ‘lower’ dating proposed by modern re-evaluations, Ithaca’s predictions are on average only 5 years off the newly proposed ground truths.\"; F43 (B4, Main, 'Contributing to historical debates' ¶4): \"Ithaca’s average predicted date for all of these decrees is 421 bc\"; F44 (B1, ¶3 s4): \"Historians have already used the tool to reevaluate significant periods in Greek history.\"", "receiving_text": "—", "translation": "untranslated: composed as received", "archive_claims": [] }, { "lineage": "the stated limits: an assistive tool, plausible suggestions, the tradition in the data", "field_claims": [ "F45", "F46", "F47", "F48", "F49" ], "received": "F45 (B4, Abstract s4–5): \"Ithaca is designed to assist and expand the historian’s workflow. The architecture of Ithaca focuses on collaboration, decision support and interpretability.\"; F46 (B4, Methods, 'Data circularity'): \"Rather, Ithaca is an assistive tool aiming to improve on and facilitate a scholarly process of induction, model uncertainty and propose possible solutions for the scholar to consider.\"; F47 (B4, Methods, 'Data circularity'): \"none of Ithaca’s proposed restorations are assumed to be objectively certain—instead, they are presented as plausible suggestions.\"; F48 (B4, Methods, 'Contributing to historical debates' (the decrees)): \"The fact that Ithaca was trained on the largest available dataset of Greek epigraphic texts makes it possible to challenge or overcome individual biases or, indeed, errors in the existing academic tradition, notwithstanding the fact that the dataset in question is originally based on the accumulated academic tradition.\"; F49 (B4, Methods, 'Training on epigraphic tasks'): \"The limitations of discretizing and amalgamating date ranges of different levels of precision based on past scholarship have been noted\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-17", "archive_claims": [ "RF-17" ] }, { "lineage": "access and licences: the interface, the code, the weights, the article", "field_claims": [ "F51", "F50", "F52", "F53", "F54", "F55" ], "received": "F51 (B4, Conclusions ¶1): \"our interdisciplinary team created an open-source and publicly available interface ( https://ithaca.deepmind.com ), enabling historians to use Ithaca for their personal research\"; F50 (B3, title): \"Predicting the Past | Contextualising, restoring, and attributing ancient texts\"; F52 (B1, ¶4): \"we partnered with Google Cloud and Google Arts & Culture to launch a free interactive version of Ithaca . And to aid further research, we have also open sourced our code, the pretrained model, and an interactive Colaboratory notebook.\"; F53 (B1, Notes): \"The Ithaca web interface was developed by Justin Grayston, Benjamin Maynard, and Ricardo Cardenas from Google Cloud.\"; F54 (B4, Code availability): \"Ithaca’s training and inference source code is available at GitHub ( https://github.com/deepmind/ithaca ) under Apache License 2.0, along with the trained weights, licensed under Creative Commons Attribution-ShareAlike 4.0 International.\"; F55 (B4, Rights and permissions): \"Open Access This article is licensed under a Creative Commons Attribution 4.0 International License\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-03, RF-22, RF-30, RF-23", "archive_claims": [ "RF-03", "RF-22", "RF-30", "RF-23" ] }, { "lineage": "genealogy and successors: Pythia before, other languages and Aeneas after", "field_claims": [ "F56", "F57", "F58", "F59", "F60" ], "received": "F56 (B4, Methods, 'Previous work' ¶2): \"The closest work to Ithaca is our 2019 research on ancient text restoration: Pythia\"; F57 (B5, Google DeepMind §Others, Archaeology s3): \"The work builds on another text analysis network that DeepMind released in 2019, named Pythia.\"; F58 (B4, Conclusions ¶1): \"Ithaca is to our knowledge the first epigraphic restoration and attribution model of its kind.\"; F59 (B1, 'Contributing to historical debates' ¶4): \"we are currently working on versions of Ithaca trained on other ancient languages and historians can already use their datasets in the current architecture to study other ancient writing systems, from Akkadian to Demotic and Hebrew to Mayan\"; F60 (B5, Google DeepMind §Others, Archaeology s7): \"The team is working on extending the model to other ancient languages, including Demotic , Akkadian , Hebrew , and Mayan .\"", "receiving_text": "—", "translation": "untranslated: composed as received beside RF-28, CAP-1-16", "archive_claims": [ "RF-28", "CAP-1-16" ] }, { "lineage": "scope: other ancient texts and disciplines", "field_claims": [ "F61" ], "received": "F61 (B4, Conclusions ¶2): \"the methods introduced in this research apply to all disciplines dealing with ancient text (papyrology, numismatics, codicology), to any language (ancient or modern)\"", "receiving_text": "—", "translation": "untranslated: composed as received", "archive_claims": [] } ], "receiving_texts": "equal-terms case: every field claim composed as received, beside the archive claims it stands with" }, "measure": { "rule": "v0.9 §4.10: KO_B sentences surviving in the KO at similarity ≥ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence)", "v0.9": { "surviving": 0, "of": 14, "positions": [ null, null, null, null, null, null, null, null, null, null, null, null, null, null ] }, "opening_run": { "rule": "v0.9 §5.10 (equal-terms case): the field-only KO's sentences as the KO's opening run in their own order", "breach": false }, "blocks": { "rule": "v0.9 §4.1 (equal-terms and mixed cases): the KO does not open with the field as a block or close with the archive as one; sentences typed by the ledgers of the claims they cite (F field-only, A archive-only, M both)", "sequence": "MFFMMMFMMFFMMFAAFMMMMFMAMMAMMA", "field_only_head": 0, "archive_only_tail": 1, "breach": false }, "max_similarity": [ 0.385, 0.29, 0.451, 0.314, 0.27, 0.49, 0.386, 0.417, 0.202, 0.414, 0.35, 0.486, 0.395, 0.278 ], "ko_sentence_1_max_similarity_to_KO_B": 0.385 }, "homonyms": { "names": [ "Homer's Ithaca, the island of the Odyssey, after which the model is named (F8, F9, F10; Wikipedia's disambiguation, F62)", "Ithaca (island), Greece, 'possibly Homer's Ithaca' (W7, the disambiguation page)", "Ithaca, New York, the city, home of Cornell University (W7); the place of publication in the archive's citations of Cornell University Press (#123, #134, #550, #605, #666, #752, #763, #816, #1270), read and rejected", "Cavafy's poem 'Ithaca' and the archive's own figure, 'Your words will bear you to Ithaca' (the Epistle to the Human Diaspora: #43, #85, #331, #402, #404, #405, #406, #1138, #1187, #1197, #1252, #1277, #1278), the Odyssean homecoming as the archive's 'survival promise', read and rejected: the homonym's, never the model's", "the other referents the disambiguation page lists (places in Australia and the United States, schools, a film, albums, a band, Ithaca Energy, the Ithaca Gun Company, an asteroid and others) (W6, W7)" ], "basis": "The model's makers name it after Homer's island (F8, F9) and the encyclopedia's bare 'Ithaca' is a disambiguation page that does not list it (F62); the operator added 'DeepMind' to the address because the bare name is a homonym (CAP-1-09, RF-29), and the registry reads that the disambiguator selects the model from the island and the city and that the composition is of the model throughout (CAP-1-12). T composes no homonym. The homonyms are distinct entities, never senses of the name.", "distinct_entities": [ "Aeneas, DeepMind's later model for Latin inscriptions (the census's relation 'neighbour', RF-28; the interface site's pair, F5; its own row in this batch, datasets/negative-of-the-negative/v2/rows/aeneas.json when seated); the aeneas census source states that Ithaca was upgraded in July 2025 to be powered by Aeneas (CAP-1-16 cites it as 'the record')", "Pythia, DeepMind's 2019 restoration model, its predecessor (F56, F57, F9)", "PHI, the Packard Humanities Institute's Searchable Greek Inscriptions, and I.PHI, its processed dataset (F18–F21)", "Predicting the Past, the interface that serves Ithaca and Aeneas (F5, F6, F50)", "Google DeepMind, the operator (B5's article)", "the Vesuvius Challenge, set beside Ithaca in #1443 (D1443-06); Enoch, the Dead Sea Scrolls dating model, set beside it in the registry's analysis (CAP-1-08) and in the census's worked question (RF-35)" ], "substitutes": { "T1": [] }, "register": { "address_1": { "slug": "ithaca-deepmind-aio-20261031", "obs_id": "OBS-5e8377899856" } } }, "notes": [ "One address, Google AI Overview (expanded), signed out, incognito, 2026-10-31 (received 00:04 EDT): ithaca-deepmind-aio-20261031 (OBS-5e8377899856, ADDR-665778bf9317, AXN:0B36.CAPTURE.🟢🌗🏠📦🕗⭐), four cards, all the entity's. A NEW address; the census record was built the day before and no card cites it (CAP-1-10).", "The homonyms (BRIEF-NEW-ROW-FIELD §5): the operator added 'DeepMind' to a bare name that is Homer's island, the Greek island, the city of Ithaca, New York and more (F62; CAP-1-09). T composes the target throughout (CAP-1-12) and no homonym; the field B is the four cards and Wikipedia. The archive's own uses of 'Ithaca' are the homonyms' (the Epistle's 'Your words will bear you to Ithaca'; Cornell University Press citations), read and rejected (traversal/ithaca/reading.json).", "Fetching: every reachable page fetched with curl and the census's agent string 'alexanarch-seating/1.0 (https://www.alexanarch.org)' on 2026-10-31 between 04:39 and 04:42 UTC, raw and text with sha256 in field.json. B3 (Predicting the Past) is byte-identical to the census's ithaca--home; B4 (Nature) has text identical to the census's ithaca--nature, its raw HTML differing only in per-request tokens; each carried once, as a card.", "Unreachable: B2, the GitHub repository google-deepmind/ithaca: HTTP 403 from the build environment's GitHub gate (a JSON notice that GitHub access is not enabled for this session), saved as B2-403.html; not retried with another agent string or route (no gh, no API, no connector). Its snippet stands as its text (F4). Wikipedia: two API search calls HTTP 429, one raw title HTTP 429 then 404 on a single retry with the same agent.", "Card snippets: B1's snippet ('Core: Ithaca deep neural network (restores ancient Greek inscriptions, …). Architecture: Sparse multi-he...') is a summary, not the post's words; B4's ('Model: Ithaca, a deep neural network trained on ancient Greek inscriptions (7th century BC to 5th century AD) from the I.PHI dataset (78,608 texts), simultaneou...') summarizes the paper's Main; B3's is the page's title tail and meta description, verbatim.", "Census fields that the field's sources show out of date or incomplete (report; not fixed): license 'not read' (RF-23), while the census's own saved copy of the paper states the code under Apache License 2.0, the trained weights under CC BY-SA 4.0 and the article under CC BY 4.0 (F54, F55); coverage.period 'as its training corpus (not stated in the abstract)' (RF-13) and count 'not stated in the saved abstract' (RF-14), while the same saved paper states 'between the seventh century bc and the fifth century ad' and I.PHI 'containing 78,608 inscriptions' (F17, F19), outside the abstract; unchecked 'the training corpus' (RF-30), which the paper describes (PHI, 178,551 texts, 84 regions; F18, F20). url 'https://ithaca.deepmind.com/' (RF-03) now redirects to https://predictingthepast.com/ (the census's own manifest records the final URL), where the interface is served with Aeneas (F5, F50).", "Capture findings the field ledger shows wrong or loosely located (report; not fixed): CAP-1-08 ('a machine dates an inscription to within thirty years and no closer') against F34, the paper's average distance of 29.3 years 'with a median distance of only 3 years': thirty years is the mean distance, not a floor. CAP-1-04 ('FIGURES BEYOND THE RECORD … come from the GitHub and Nature cards; the record did not read the licence'): the regions, the architecture, PHI and the Apache licence are all in the paper the census saved (F18, F24, F20, F54); beyond the record's fields, not beyond its saved source. CAP-1-16 cites 'the record: upgraded in July 2025 to be powered by Aeneas': the Ithaca census record does not say so; the statement is in the aeneas census source (aeneas--blog.txt L337). The registry entry's per_dim.read.on is '2026-10-29', two days before the capture (received 2026-10-31); per_dim.read.notes' 'L84, L191' are line numbers of the census's ithaca--nature.txt.", "Neighbour rows (BRIEF-NEW-ROW-FIELD §6): datasets/negative-of-the-negative/v2/rows/book-of-revelation.json is the text the field is organized around; Ithaca bears on it only negatively in the census's terms (RF-20 'anything on the New Testament' among what it cannot answer; RF-27; RF-35). datasets/negative-of-the-negative/v2/rows/sappho-31.json: no bearing found; that row does not name Oxyrhynchus, and Ithaca's only reach toward papyri is the authors' claim that the methods apply to papyrology (F61) and the census's reading that it bears 'only as machine dating moves toward literary and papyrus evidence' (RF-21). Sibling rows of this batch: aeneas (the census's neighbour, RF-28) and enoch-dss-dating (the registry's comparison, CAP-1-08; the worked question, RF-35). None edited.", "Archive admitted: two deposits of 81 candidate deposits read (#1443 at D, R and O-direct, 6 claims; #1406 by the hop at O, 2 claims), with the census record (35 claims) and the Capture Registry's entry (17 claims) ledgered as the archive's records of the entity. The census's rf_intersection and correction are the archive's reading, composed as the census record's and never as the field's (RF-21, RF-32).", "The KO closes with one archive-only sentence (the registry's analysis); its archive-only sentences are the census's question (¶14) and the registry's account of the composition (¶15), senses only the archive's records state. Measure: 0 of 14 KO_B sentences survive at ≥ 0.95; KO sentence 1's best similarity to a KO_B sentence 0.385." ], "composition_v09": { "source": "this row (composed directly; no rebuild/negative-of-the-negative/v09/ithaca.json)", "composer": "Claude (Opus 5.5), subagent, 2026-10-31", "defining_material_read": [ { "what": "Revelation First census record, row", "file": "/home/claude/fleet/revelationfirst-com/data/rf-census/census.json (rows[slug == 'ithaca'], organizing_question, correction, intersections, evidence_kinds, schema_note, pending, unread_sources)", "read": "whole" }, { "what": "Revelation First worked questions", "file": "/home/claude/fleet/revelationfirst-com/data/rf-census/questions.json", "read": "the question that names the entity ('machine-dating')" }, { "what": "Revelation First census record, rendered", "file": "/home/claude/fleet/revelationfirst-com/aggregators/ithaca/index.html", "read": "text, for agreement with the row" }, { "what": "census sources", "file": "/home/claude/fleet/revelationfirst-com/data/rf-census/sources/ithaca--nature.*, ithaca--home.*, manifest.json; rows/ithaca.py", "read": "whole" }, { "what": "the Capture Registry entry", "file": "data/EA-WG-CAPTURES-01.json (entries[slug == 'ithaca-deepmind-aio-20261031'])", "read": "whole" }, { "what": "#1443 The Photograph of the Handwritten Original", "file": "data/texts/AXN-05CC-text.md", "read": "front matter, Parts I–II, §2.3 whole, Part X, the reference list R·2" }, { "what": "#1406 Differential Register Prioritization", "file": "data/texts/AXN-058F-text.md", "read": "description, methodology, falsification conditions, abstract, §0–§1" }, { "what": "#202 Revelation First ≠ Revelation Early", "file": "data/texts/AXN-0349-text.md", "read": "searched for machine dating; no bearing" }, { "what": "the field's saved texts", "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B1.txt, B3-meta.txt, B4.txt, B5.txt, W7.txt", "read": "B1, B3-meta whole; B4 abstract, Main, Methods (previous work, I.PHI, torso, training, data circularity, the decrees), data and code availability, rights, acknowledgements, author information; B5 the 'Archaeology' section; W7 the disambiguation" } ], "cuts": [], "notes": "equal case; KO 30 sentences, all 62 field and 60 archive claims cited; P body 319 words; KO_B 14 sentences, 0 surviving at ≥ 0.95" }, "field_by_origin": { "cards": [ "B1", "B2", "B3", "B4" ], "homonym_cards": [], "own_pages": [], "wikipedia": [ "B5" ], "census_third_party": [], "unreachable": [ "B2" ], "title_only": [], "own_pages_note": "The census's sources for the entity are two, both surfaced by T as cards and carried once, as cards: ithaca--nature (the paper, the census's evidence source, basis 'operator') = B4, text identical; ithaca--home (https://ithaca.deepmind.com/, which redirects to https://predictingthepast.com/), which the census lists in unread_sources (http 200, a JavaScript application) = B3, byte-identical. No own page outside T's rail remains to admit. The census read no third-party source for the entity.", "wikipedia_state": "no article: raw wikitext HTTP 404 for 'Ithaca (neural network)', 'Ithaca (DeepMind)', 'Ithaca (artificial intelligence)', 'Ithaca (machine learning)', 'Ithaca (model)' (rendered 'Ithaca (neural network)' HTTP 404); the bare 'Ithaca' is a disambiguation page (HTTP 200) that lists Homer's Ithaca, the island, the city of Ithaca, New York and other referents and does not list the model; a full-text search finds the model in 'Google DeepMind' (revision 1380412362), section 'Others' › 'Archaeology', ledgered as B5 (F7, F10, F35, F40, F41, F57, F60), with the disambiguation's opening (F62); 'Epigraphy' does not mention it. Two API calls returned HTTP 429 and were not repeated.", "sources": [ { "id": "B1", "origin": "card", "role": "Overview card 1", "card": "Google DeepMind — Predicting the past with Ithaca — Google DeepMind", "url": "https://deepmind.google/discover/blog/predicting-the-past-with-ithaca/", "kind": "Overview card (the operator's post of 9 March 2022, by Yannis Assael, Thea Sommerschield, Brendan Shillingford and Nando de Freitas)", "files": { "B1.html": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B1.html", "sha256": "832d3d6cfd97b40ad266c0fe64f28f9e9e3eabab66f1d73a47403462f5a5ff26" }, "B1.txt": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B1.txt", "sha256": "6de295b7663778e51d28c7260a573054deeb3011e09049ba59889c08f10f5292" } }, "fetched": { "at": "2026-10-31T04:39:12Z", "http": 200, "url": "https://deepmind.google/discover/blog/predicting-the-past-with-ithaca/", "final_url": "https://deepmind.google/blog/predicting-the-past-with-ithaca/", "raw": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B1.html", "raw_sha256": "832d3d6cfd97b40ad266c0fe64f28f9e9e3eabab66f1d73a47403462f5a5ff26", "agent": "alexanarch-seating/1.0 (https://www.alexanarch.org)", "text": [ "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B1.txt" ] }, "names_operator": "yes ('DeepMind'; the authors Yannis Assael, Thea Sommerschield, Brendan Shillingford, Nando de Freitas; the partners Ca' Foscari, Oxford, AUEB; the web interface's developers at Google Cloud)", "note": "URL by WebSearch on the card's title and site (deepmind.google/discover/blog/…, which redirects to deepmind.google/blog/…); the page's is the card's title. Fetched once with the census's agent string. Not a census source.", "surfaced_by_T": [ { "slug": "ithaca-deepmind-aio-20261031", "card": 1 } ], "card_title": "Predicting the past with Ithaca — Google DeepMind", "card_snippet": "Core: Ithaca deep neural network (restores ancient Greek inscriptions, identifies geographic origins, establishes creation dates). Architecture: Sparse multi-he...", "card_note": "the snippet ('Core: … Architecture: Sparse multi-he...') is a summary of the page, not its words", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" }, { "id": "B2", "origin": "card", "role": "Overview card 2", "card": "GitHub — GitHub - google-deepmind/ithaca: Restoring and attributing ancient texts using deep neural networks · GitHub", "url": "https://github.com/google-deepmind/ithaca", "kind": "Overview card (the code repository)", "files": { "B2-403.html": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B2-403.html", "sha256": "b3e03aaaff81d68730d67392a135ac3fdfdf022880c66632ab188d2fe084cda3" } }, "fetched": { "at": "2026-10-31T04:39:12Z", "http": 403, "url": "https://github.com/google-deepmind/ithaca", "raw": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B2-403.html", "raw_sha256": "b3e03aaaff81d68730d67392a135ac3fdfdf022880c66632ab188d2fe084cda3", "agent": "alexanarch-seating/1.0 (https://www.alexanarch.org)", "text": [], "result": "unreachable: HTTP 403 from the build environment's GitHub gate (a JSON message that GitHub access to the repository is not enabled for this session), saved as B2-403.html; not retried with another agent string or by any other route (no gh, no API, no connector)" }, "names_operator": "yes, in the snippet ('developed by DeepMind and academic partners')", "note": "unreachable (BRIEF-NEW-ROW-FIELD §4): the URL from the card's title (repository google-deepmind/ithaca; the paper's Code availability gives github.com/deepmind/ithaca, the organisation's earlier name). The card's snippet stands as its only text (F4).", "surfaced_by_T": [ { "slug": "ithaca-deepmind-aio-20261031", "card": 2 } ], "card_title": "GitHub - google-deepmind/ithaca: Restoring and attributing ancient texts using deep neural networks · GitHub", "card_snippet": "Overview: Ithaca is a deep neural network developed by DeepMind and academic partners for restoring, geographically attributing, and chronologically dating anci...", "card_note": null, "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" }, { "id": "B3", "origin": "card", "role": "Overview card 3 (also the census's own page ithaca--home, which redirects here)", "card": "Predicting the Past — Predicting the Past", "url": "https://predictingthepast.com/", "kind": "Overview card (the public interface for Ithaca and Aeneas; ithaca.deepmind.com redirects to it)", "files": { "B3.html": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B3.html", "sha256": "08a11cf87711fe870d18b83f8adc34ec85b625bad14b3b646dab7052d6d393a9" }, "B3.txt": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B3.txt", "sha256": "06b41db902580a817c82f9842d473ba88f1965be698259d506dedf7a0c88e8f7" }, "B3-meta.txt": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B3-meta.txt", "sha256": "3f6895b72c46ded7bddb1adf2ba58a8838c1dce03f889e34d7f5e8d357529c74" } }, "fetched": { "at": "2026-10-31T04:39:13Z", "http": 200, "url": "https://predictingthepast.com/", "raw": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B3.html", "raw_sha256": "08a11cf87711fe870d18b83f8adc34ec85b625bad14b3b646dab7052d6d393a9", "agent": "alexanarch-seating/1.0 (https://www.alexanarch.org)", "census_copy": "ithaca--home (url https://ithaca.deepmind.com/, final https://predictingthepast.com/)", "census_sha256": "08a11cf87711fe870d18b83f8adc34ec85b625bad14b3b646dab7052d6d393a9", "agrees": true, "text": [ "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B3.txt", "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B3-meta.txt" ] }, "names_operator": "no (the page's words name Ithaca and Aeneas, not their maker)", "note": "The page is a JavaScript application: its extracted text (B3.txt, identical to the census's ithaca--home.txt) is the title and a browser notice; its words are the <title> and the description in its meta tags, extracted verbatim to B3-meta.txt (the card's snippet is title tail plus the description). Byte-identical to the census's saved copy of its own page ithaca--home, which the census lists among unread_sources (http 200; a JavaScript application); carried once, as a card. The application's script bundle was not fetched.", "surfaced_by_T": [ { "slug": "ithaca-deepmind-aio-20261031", "card": 3 } ], "card_title": "Predicting the Past", "card_snippet": "Contextualising, restoring, and attributing ancient texts Ithaca and Aeneas are generative AI models designed to assist historians in the study of ancient Greek...", "card_note": null, "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" }, { "id": "B4", "origin": "card", "role": "Overview card 4 (also the census's one source, ithaca--nature)", "card": "Nature — Restoring and attributing ancient texts using deep neural networks | Nature", "url": "https://www.nature.com/articles/s41586-022-04448-z", "kind": "Overview card (the paper: Assael, Sommerschield et al., Nature 603, 280–283, 9 March 2022, open access)", "files": { "B4.html": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B4.html", "sha256": "7bab9e1b3786b154379e2f2cbf42bda24ffe81228e99404bcaea5af143217fcf" }, "B4.txt": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B4.txt", "sha256": "d10c3328114f85d4738563c21ee3136e6942603470b89e88ace9c0ee7b0dc9e1" }, "B4-census.html": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B4-census.html", "sha256": "d2b4df37a93252c3d44127e213464c4bed2ab3a26dbffde58a648645bcad7aed" } }, "fetched": { "at": "2026-10-31T04:39:15Z", "http": 200, "url": "https://www.nature.com/articles/s41586-022-04448-z", "final_url": "https://www.nature.com/articles/s41586-022-04448-z?error=cookies_not_supported&code=bfb45777-eddd-4f7d-9a38-099aec6b358b", "raw": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B4.html", "raw_sha256": "7bab9e1b3786b154379e2f2cbf42bda24ffe81228e99404bcaea5af143217fcf", "agent": "alexanarch-seating/1.0 (https://www.alexanarch.org)", "census_copy": "ithaca--nature", "census_sha256": "d2b4df37a93252c3d44127e213464c4bed2ab3a26dbffde58a648645bcad7aed", "census_fetched": "2026-10-31T02:52:07Z", "agrees": "text identical (B4.txt equals the census's ithaca--nature.txt byte for byte); the raw HTML differs only in per-request tokens (ad-call cache-busters, a CSRF value, reference-anchor ids)", "text": [ "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B4.txt" ] }, "names_operator": "yes (authors Yannis Assael, Thea Sommerschield, Brendan Shillingford, Mahyar Bordbar, John Pavlopoulos, Marita Chatzipanagiotou, Ion Androutsopoulos, Jonathan Prag, Nando de Freitas; affiliations DeepMind, Ca' Foscari, Harvard's Center for Hellenic Studies, AUEB, Oxford)", "note": "The census's only source for the row (basis 'operator' for its four evidence quotes), surfaced by T as card 4; refetched with the census's agent string and carried once, as a card, with the census's copy saved beside it (B4-census.html, gunzipped from ithaca--nature.html.gz).", "surfaced_by_T": [ { "slug": "ithaca-deepmind-aio-20261031", "card": 4 } ], "card_title": "Restoring and attributing ancient texts using deep neural networks | Nature", "card_snippet": "Model: Ithaca, a deep neural network trained on ancient Greek inscriptions (7th century BC to 5th century AD) from the I.PHI dataset (78,608 texts), simultaneou...", "card_note": null, "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" }, { "id": "B5", "origin": "wikipedia", "role": "Wikipedia (no article on the entity; the section that describes it)", "card": "Wikipedia — Google DeepMind §Others, 'Archaeology' (no article 'Ithaca (neural network)' or similar; 'Ithaca' a disambiguation page that does not list the model)", "url": "https://en.wikipedia.org/wiki/Google_DeepMind", "kind": "encyclopedia section (revision 1380412362) and the disambiguation page of the bare name", "files": { "B5.html": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B5.html", "sha256": "4e7bfba0afa24409ab25ac687952cd2e16958d318b815e30cb45f07c1ec9a73f" }, "B5.txt": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B5.txt", "sha256": "e0065367800a53a04e8c97d94eb089590cf83a02edd5f447159ed6358eec11c8" }, "W7.txt": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W7.txt", "sha256": "91ad081371380a52d1901b8e915b362e19b404b5bdfa6ce4f20ff8bdd12f21a5" }, "W1.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W1.raw", "sha256": "11567ee84c27a1249fbe65bff7a643b8ed0fe1bdc419109d83e7108c4cb6d227" }, "W2.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W2.raw", "sha256": "3732a8dc9c5a26a4246ec2e6f9107107e740d0f6f5a8b3645e0fd055ed277e8e" }, "W3.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W3.raw", "sha256": "edc480d3fce87911b4cee00b27527b08635403b219c839d18954bf8bac3d5794" }, "W4.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W4.raw", "sha256": "30fe3bb6eccd18d5339bb8caf4619e358f975e1cbaf76c2b3ad4409379006ebe" }, "W5.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W5.raw", "sha256": "20d07d161c01ef8385194b7a85730c73bd55d88bfefffcc031d166d1f5fc933f" }, "W6.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W6.raw", "sha256": "1ebc4422bd2be84f26549f2468098fcc7ad986cb4ccfe27465b2ad65e6756226" }, "W7.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W7.raw", "sha256": "97c959ea17bdaaa7f42833a0b80e1f5e447970b67cdfa1ac2a26a86efb11bb40" }, "W8.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W8.raw", "sha256": "d700293d0bd047ac312802ead90ab6a01168e8f6c1bcd63daa5d69e3855cd598" }, "W9.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W9.raw", "sha256": "133497d3b16d94d33e372b86f442d6cc9cf5299a15e996b1182603807f9abb0a" }, "W10.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W10.raw", "sha256": "0b776434b568a88e521e24556f10e9053f2de834e735c8cf9e8b6cc0ab8dfdfc" }, "W11.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W11.raw", "sha256": "f2610489fa927fd4f4c017d5067267fae04be4a13f4a556cd7bf869d85c6cb22" }, "W12.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W12.raw", "sha256": "c4c643ec29a42b7f9f1b0ee31f653f2f147ff1c9ed6fe671b26ca4579e54acb5" }, "W13.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W13.raw", "sha256": "4e7bfba0afa24409ab25ac687952cd2e16958d318b815e30cb45f07c1ec9a73f" }, "W14.raw": { "path": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W14.raw", "sha256": "bf0bb48961467553aae2008971ff67fb64cc32aae0efc6a95b6fe71e57f2eae3" } }, "fetched": { "responses": [ { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W1.raw", "url": "https://en.wikipedia.org/w/index.php?title=Ithaca_(neural_network)&action=raw", "http": 404, "at": "2026-10-31T04:41:04Z", "sha256": "11567ee84c27a1249fbe65bff7a643b8ed0fe1bdc419109d83e7108c4cb6d227" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W2.raw", "url": "https://en.wikipedia.org/w/index.php?title=Ithaca_(DeepMind)&action=raw", "http": 404, "at": "2026-10-31T04:41:04Z", "sha256": "3732a8dc9c5a26a4246ec2e6f9107107e740d0f6f5a8b3645e0fd055ed277e8e" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W3.raw", "url": "https://en.wikipedia.org/w/index.php?title=Ithaca_(artificial_intelligence)&action=raw", "http": 429, "at": "2026-10-31T04:41:04Z", "sha256": "edc480d3fce87911b4cee00b27527b08635403b219c839d18954bf8bac3d5794" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W4.raw", "url": "https://en.wikipedia.org/w/index.php?title=Ithaca_(machine_learning)&action=raw", "http": 404, "at": "2026-10-31T04:41:05Z", "sha256": "30fe3bb6eccd18d5339bb8caf4619e358f975e1cbaf76c2b3ad4409379006ebe" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W5.raw", "url": "https://en.wikipedia.org/w/index.php?title=Ithaca_(model)&action=raw", "http": 404, "at": "2026-10-31T04:41:05Z", "sha256": "20d07d161c01ef8385194b7a85730c73bd55d88bfefffcc031d166d1f5fc933f" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W6.raw", "url": "https://en.wikipedia.org/w/index.php?title=Ithaca&action=raw", "http": 200, "at": "2026-10-31T04:41:05Z", "sha256": "1ebc4422bd2be84f26549f2468098fcc7ad986cb4ccfe27465b2ad65e6756226" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W7.raw", "url": "https://en.wikipedia.org/wiki/Ithaca", "http": 200, "at": "2026-10-31T04:41:05Z", "sha256": "97c959ea17bdaaa7f42833a0b80e1f5e447970b67cdfa1ac2a26a86efb11bb40" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W8.raw", "url": "https://en.wikipedia.org/wiki/Ithaca_(neural_network)", "http": 404, "at": "2026-10-31T04:41:06Z", "sha256": "d700293d0bd047ac312802ead90ab6a01168e8f6c1bcd63daa5d69e3855cd598" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W9.raw", "url": "https://en.wikipedia.org/w/index.php?title=Google_DeepMind&action=raw", "http": 200, "at": "2026-10-31T04:41:15Z", "sha256": "133497d3b16d94d33e372b86f442d6cc9cf5299a15e996b1182603807f9abb0a" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W10.raw", "url": "https://en.wikipedia.org/w/index.php?title=Epigraphy&action=raw", "http": 200, "at": "2026-10-31T04:41:16Z", "sha256": "0b776434b568a88e521e24556f10e9053f2de834e735c8cf9e8b6cc0ab8dfdfc" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W11.raw", "url": "https://en.wikipedia.org/w/api.php?action=query&list=search&srsearch=Ithaca+DeepMind+inscriptions&format=json&srlimit=20", "http": 429, "at": "2026-10-31T04:41:16Z", "sha256": "f2610489fa927fd4f4c017d5067267fae04be4a13f4a556cd7bf869d85c6cb22" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W12.raw", "url": "https://en.wikipedia.org/w/index.php?title=Ithaca_(artificial_intelligence)&action=raw", "http": 404, "at": "2026-10-31T04:41:16Z", "sha256": "c4c643ec29a42b7f9f1b0ee31f653f2f147ff1c9ed6fe671b26ca4579e54acb5" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W13.raw", "url": "https://en.wikipedia.org/wiki/Google_DeepMind", "http": 200, "at": "2026-10-31T04:41:24Z", "sha256": "4e7bfba0afa24409ab25ac687952cd2e16958d318b815e30cb45f07c1ec9a73f" }, { "file": "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W14.raw", "url": "https://en.wikipedia.org/w/index.php?search=%22Ithaca%22+DeepMind+inscriptions&title=Special:Search&profile=advanced&fulltext=1&ns0=1", "http": 200, "at": "2026-10-31T04:41:44Z", "sha256": "bf0bb48961467553aae2008971ff67fb64cc32aae0efc6a95b6fe71e57f2eae3" } ], "text": [ "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/B5.txt", "datasets/negative-of-the-negative/v2/intake/ithaca-20261031/W7.txt" ], "agent": "alexanarch-seating/1.0 (https://www.alexanarch.org)" }, "names_operator": "yes ('Google'; 'DeepMind'; the cited post's authors in the wikitext)", "note": "No article for the entity: raw wikitext for 'Ithaca (neural network)', 'Ithaca (DeepMind)', 'Ithaca (artificial intelligence)', 'Ithaca (machine learning)', 'Ithaca (model)' HTTP 404 (one 429 rate-limit response for 'Ithaca (artificial intelligence)', retried once with the same agent: 404). 'Ithaca' is a disambiguation page (W6 wikitext, W7 rendered) that does not list the model. A full-text search (W14) finds the model only in 'Google DeepMind', section 'Others' › 'Archaeology' (W9 wikitext, W13 = B5.html rendered), which is ledgered; two API search calls returned HTTP 429 (W11 and an earlier unsaved one) and were not repeated. 'Epigraphy' (W10) does not mention it.", "surfaced_by_T": [], "card_title": null, "card_snippet": null, "card_note": "not surfaced by T; admitted on the operator's ruling", "names_archive": "no (none of the Crimson Hexagonal Archive, Alexanarch, Lee Sharks, revelationfirst, machinemediation, EA-WG-CAPTURES-01 in the saved text, card title and snippet)" } ] } }