Skip to content

World's first open-source enterprise world model.

What prior work taught us (#742)

WaylandYang2f9bdf7cbe70
diff --git a/docs/design/README.md b/docs/design/README.md
index 03b104da..4e8b24e1 100644
--- a/docs/design/README.md
+++ b/docs/design/README.md
@@ -50,7 +50,7 @@ below marks these records from this note; their own status lines predate it and
`current`: every decision in the record holds. `partly superseded (by NNNN)`: a later record, or
the note above, overturned some of its decisions and the rest hold; the record's own revision notes
say which. `superseded (by NNNN)`: none of its decisions hold. `proposed`: the direction is accepted
-and the record is not yet on `dev` (0034 has no code; 0043 sits in PR #699; 0044 and 0045 sit in
+and the record is not yet on `dev` (0034 has no code; 0043 was in PR #699, closed on 2026-09-17 to re-land on the open graph after alignment; 0044 and 0045 sit in
PRs #710 and #724, and 0044's cut 1 has landed ahead of the record). The record's own status line
is the source of truth for what is built; this table only adds whether a later record has overtaken
it.
@@ -101,10 +101,13 @@ it.
| 0040 | 2026-09-13 | [A chunk says where its words came from](../decisions/0040-a-chunk-says-where-its-words-came-from.md) | sources | current |
| 0041 | 2026-09-13 | [A name is a claim about an entity](../decisions/0041-a-name-is-a-claim-about-an-entity.md) | identity | current |
| 0042 | 2026-09-13 | [The chat loop is a runner with hooks](../decisions/0042-the-chat-loop-is-a-runner-with-hooks.md) | chat-and-mcp | current |
-| 0043 | 2026-09-14 | Every review queue is governed (in PR #699, not yet in this directory) | governance | proposed |
+| 0043 | 2026-09-14 | Every review queue is governed (was PR #699, closed 2026-09-17; re-lands after 0044 cut 2) | governance | proposed |
| 0044 | 2026-09-16 | [The ontology is a view over what documents say](../decisions/0044-the-ontology-is-a-view-over-what-documents-say.md) | ontology | proposed |
| 0045 | 2026-09-16 | [A time mention is resolved against its document](../decisions/0045-a-time-mention-is-resolved-against-its-document.md) | time | proposed |
+[prior-work.md](prior-work.md) is not a domain: it places each layer in the literature it stands on
+and lists the pitfalls taken from it, with what is still open.
+
0016 is a schedule and a convention rather than a design; its convention (the PR that implements a
record updates its status line) lives in the [decisions README](../decisions/README.md), and its
lines are tracked in the domain files they belong to. [pipeline.md](../pipeline.md) still describes
diff --git a/docs/design/extraction.md b/docs/design/extraction.md
index ec3defd3..97351030 100644
--- a/docs/design/extraction.md
+++ b/docs/design/extraction.md
@@ -23,6 +23,24 @@ and carry their own quote because a numbered contract mixed the ids up on dense
10 to 14% with ids, 2 to 7% with names) [#731]. Parsing is per item: a malformed item is counted and
a truncated reply is repaired to its last complete item.
+**What is a thing, what is a statement** [#743, prior-work items 1 and 6]. A thing is named when its
+name is a proper name or a fixed term, one that means the same thing in any document (a person, an
+organization, a product, a place, a document, a law, an event; a disease, a drug, an industry, a
+product category, an indicator); it is described when it is a role or a generic phrase whose
+referent the passage decides ("the company", "patients", "各部门"), however particular it is
+there. Named things resolve across documents by name [0041]; described things exist only in their
+document and only when a statement points at them. A phrase is the verb with the words that belong
+to it, so that subject, phrase and object read as a sentence on their own, never a bare verb cut
+from a longer verb phrase; several verbs sharing one object are one statement, one verb with several
+objects is one statement per object. The object is what the verb acts on; where, how, why, with
+what and for whom go in qualifiers under the passage's own role word, a named thing mentioned there
+is still listed and the qualifier names it, and a generic phrase that appears only there stays
+words in the qualifier. In a table, a cell is a statement about its row's thing whose phrase is
+the column heading; when the heading names a time it is `when` instead, the statement is about
+the thing the caption names and the phrase is the row label with its section path, and a unit the
+caption gives is a qualifier [#744]. These are contract rules, not server checks: the shape checks
+below look at structure, never at vocabulary.
+
**Server checks, each a drop reason in `extraction_drops`.** Every quote must occur in the chunk
(`quote_not_in_chunk`); every name in its quote (`name_not_in_text`); a time mention only when its
words occur in the statement's own quote (`time_not_in_quote`), because the model otherwise attaches
@@ -30,8 +48,11 @@ one time to unrelated sentences; every subject must be a listed thing (`unknown_
object becomes a literal value with a signal (`object_undeclared`); a string one entity already
claims in the document is not a name of another (`name_claimed_by_another`). Offsets are computed on
the server by locating the quote, never taken from the model. The remaining codes are
-`malformed_item`, `truncated_reply` and `object_missing`; the 28 codes of the typed path went with
-it [#736]. A drop is a row, never silence; the table is cleared per document when extraction starts
+`malformed_item`, `truncated_reply` and `object_missing`; a phrase that is the value itself or the
+subject's own name is kept and counted (`phrase_is_value`, `phrase_is_subject`, the two shapes a
+model writes for a table row that lost its heading or its caption; the subject and the value are
+there, so nothing is lost, #744); the 28 codes of the typed path went with it [#736]. A drop is a
+row, never silence; the table is cleared per document when extraction starts
and shown per document in the Library, apart from Review [0001, 0005].
**What it writes** (see [ledger](ledger.md)). Open statements as `facts` rows with `layer = 'open'`,
@@ -73,7 +94,9 @@ qualifiers, time words and quote span, and a nod writes an open statement [0015,
`scripts/bench/recall.mjs` (SEC filings, pharma, ai-timeline; Re-DocRED fetched by script and never
used as prompt examples) scores entity-pair recall; `judge_open.mjs` has a judge model read the
-chunk and reports stated, misworded and not stated; `identity.mjs`, `govern.mjs`, `temporal.mjs`
+chunk and reports stated, misworded and not stated, and separately whether the statement reads on
+its own without the document (`alone`, the uninformative-phrase class of [prior-work](prior-work.md)
+item 1); `identity.mjs`, `govern.mjs`, `temporal.mjs`
and the lease bench cover the other domains. Every cut reports at least three domains, two runs per
configuration, with the judge's calibration stated; no F1 against Re-DocRED, whose gold omits true
facts [0044]. Thresholds: not stated at most 2%, entity-pair recall no lower than before, prompt
diff --git a/docs/design/prior-work.md b/docs/design/prior-work.md
new file mode 100644
index 00000000..63ec183d
--- /dev/null
+++ b/docs/design/prior-work.md
@@ -0,0 +1,535 @@
+# What prior work taught us
+
+The three layers of [0044] and the time resolution of [0045] were not invented here. Each stands on
+a line of research that is fifteen to forty years old, and each of those lines has already met the
+failures we meet on our own corpora. This file places every layer in that literature and lists the
+pitfalls we took from it: who found it, what we do about it, and what is still missing. It changes
+when a design file changes for a reason the literature had already given, or when a new pitfall is
+taken. A paper is cited as [Fader 2011]; the list at the end gives venue, year and a link for every
+one, and every entry was checked against its publisher page before it was written here.
+
+## Where each layer stands
+
+**The open graph is extractive Open Information Extraction.** A statement in the document's own
+words, with the relation phrase as written and no schema, is the programme Banko et al. opened
+[Banko 2007] and the Etzioni group carried through ReVerb [Fader 2011] and OLLIE [Mausam 2012];
+ClausIE [Del Corro 2013], Stanford OpenIE [Angeli 2015], the supervised
+turn [Stanovsky 2018] and OpenIE6 [Kolluru 2020] are the same programme with better parsers, and
+three surveys cover it to 2024 [Niklaus 2018, Zhou 2022, Liu 2024]. Pei et al. draw the line we sit
+on: extractive OpenIE keeps surface forms, abstractive OpenIE may infer [Pei 2023]. For Chinese the
+reference points are CORE [Tseng 2014], ZORE [Qiu 2014], the SAOKE corpus behind Logician, whose
+annotation scheme has a class of descriptive facts that matches our described things [Sun 2018],
+and Title2Event, the largest manually annotated set of Chinese open triples [Deng 2022]. Our
+departures from that tradition are three: a language model reads the chunk under a JSON contract
+instead of a parser; every statement carries its own verbatim quote located by character offsets;
+the run is at temperature 0 with a judge that counts what the text did not say
+([extraction](extraction.md)). Two recent designs converge on the same shape from the other side:
+TRACE-KG keeps a raw and a canonical predicate, qualifiers and a justification excerpt on every
+relation [Abolhasani 2026], and Han and Lai argue for natural-language relations over a minimal
+structural backbone [Han 2026].
+
+**Qualifiers beside the edge are MinIE's attributes and Wikidata's qualifiers.** OLLIE first
+separated attribution and condition from the triple [Mausam 2012]; MinIE moved polarity, modality,
+attribution and quantity out of the relation phrase into annotations [Gashteovski 2017]; NestIE kept
+nested propositions as such [Bhutani 2016]; DisSim's split of a sentence into core and context
+linked by a rhetorical relation is the same cut made by rule [Niklaus 2019]. In the ledger the idea
+is Wikidata's qualifier on a statement [Vrandečić 2014] and the hyper-relational fact of HyperRED,
+whose 44 qualifier types are the comparison point for our role words [Chia 2022]; Text2NKG lists the
+four incompatible ways an n-ary fact is serialised [Luo 2024], and ours is the hyper-relational one.
+`statement_qualifiers` keyed by the document's role word is that model with the role word left as
+written [0037, 0044].
+
+**The ontology as a view, with alignment later, is open-KB canonicalization plus pay-as-you-go
+integration.** Canonicalizing the noun phrases and relation phrases of an open knowledge base after
+extraction is the problem Galárraga et al. posed [Galárraga 2014] and CESI [Vashishth 2018a], CUVA
+[Dash 2021], JOCL [Liu 2021] and CMVC [Shen 2022] took up; Yang and Curry survey that line and find
+relation-phrase canonicalization its least solved part [Yang 2024]. Mapping open extractions onto a
+fixed schema was done for domain relations [Soderland 2010], for KBP slots [Soderland 2013, Angeli
+2015], for DBpedia [Dutta 2015, Gashteovski 2020] and for ConceptNet [Romero 2023]. The LLM-era
+pipeline closest to [0044] is EDC, which extracts open triples, writes a definition for each
+extracted relation and then lets the model select a schema element or reject them all [Zhang 2024];
+KGGen [Mo 2025], iText2KG [Lairgi 2024], AutoSchemaKG [Bai 2025], SAC-KG [Chen 2024b], Docs2KG
+[Sun 2025] and ODKE+ [Khorshidi 2025] are the same order of operations at other scales, and Bian's
+survey names the schema-based and schema-free paradigms they sit between [Bian 2025]. The stance
+that data is stored first and semantics added as it pays for itself is the dataspaces programme
+[Franklin 2005, Halevy 2006, Madhavan 2007]; an ontology that is a view over stored data through
+mappings is ontology-based data access [Poggi 2008, Xiao 2018]. The knowledge-graph surveys treat an
+emergent schema as a normal third mode beside a prescribed one [Hogan 2021, Weikum 2021].
+
+**A kind word bound to a class is ultra-fine entity typing, decided by a judge that has been shown
+its own bias.** Choi et al. defined types as the free noun phrases a text uses, with head words as
+the supervision [Choi 2018]; our kind word is that phrase kept verbatim, and the binding is its
+mapping to the ontology's class, the step Dai and Zeng treat as its own task [Dai 2023]. Typing by
+description alone [Obeidat 2019], by enriched classes [Ouyang 2024] and top-down through the
+hierarchy [Komarlu 2024] are the three moves the binding call makes. The candidate-then-judge shape,
+with an explicit "none", is the shape of LLM ontology matching [He 2023, Hertling 2023, Babaei
+Giglou 2024, Qiang 2024, Song 2025]. Deciding with two votes and the candidate order reversed on the
+second comes from the position bias measured in LLM judges and choosers [Zheng 2023, Wang 2024,
+Pezeshkpour 2024, Zheng 2024]; agreement across samples as a confidence signal is self-consistency
+[Wang 2023a].
+
+**Time resolution is temporal tagging over a bitemporal ledger.** The words of a mention and their
+value are TimeML's TIMEX3 [Pustejovsky 2003]; computing the value against a reference time by rule
+is what HeidelTime [Strötgen 2012, Strötgen 2013] and SUTime [Chang 2012] do, and TempEval-3 measured
+how much harder the value is than the span [UzZaman 2013]. The split we made, a model that returns
+an interpretation and code that computes from it, is TimeNorm's grammar given an anchor
+[Bethard 2013], SCATE's compositional operators [Bethard 2016a] and, closest of all, the 2025 framing
+of normalization as generating a program that a deterministic library executes [Su 2025a]; the
+surveys are [Leeuwenberg 2019, Zhong 2023, Su 2025b]. Dating a document is its own task
+[Chambers 2012, Vashishth 2018b]; we transcribe the date the document states instead of classifying.
+Facts with validity intervals are YAGO2's model, which "never assigns an ill-defined time"
+[Hoffart 2013], and the temporal scoping line [Ling 2010, Wang 2011, Talukdar 2012]; the LLM-era
+temporal graphs separate observed time from valid time the same way [Rasmussen 2025, Lairgi 2025].
+The two axes of [0019] are the valid time and transaction time of bitemporal databases
+[Snodgrass 1999, Jensen 1999, Kulkarni 2012], in the words of the consensus glossary [Jensen 1998];
+the precision columns are time granularity [Bettini 2000]; `attested_to` is a now-relative value in
+the sense of Clifford et al. [Clifford 1997]. The clinical corpora are where relative time was
+studied hardest [Sun 2013, Styler 2014].
+
+**A statement with a located quote is KBP provenance and an atomic proposition.** TAC KBP required
+a document offset as provenance for every slot fill; FActScore [Min 2023] and Dense X Retrieval
+[Chen 2024a] made the atomic proposition the unit of checking and of retrieval, and the quote-then-answer
+models [Menick 2022, Huang 2024, Zhang 2025] made the verbatim quote the unit of trust. Our statement
+is that unit with its provenance attached at write time [0001, 0044].
+
+## The pitfalls we took
+
+Each item: where it was found, what we do, what is still missing. Items marked *open* are not yet
+built.
+
+### Extraction
+
+1. **Uninformative relation phrases.** ReVerb named the two error classes of open extraction,
+ incoherent phrases and uninformative ones, its example being "made a deal with" cut to "made"
+ [Fader 2011]; the fact-level re-annotation of BenchIE rejected vague relation phrases outright
+ [Lamarche 2024]; labelling chunks instead of tokens is one remedy [Dong 2023a]. Our
+ "各地区 深入 企业" is ReVerb's second class. We keep the whole sentence as the quote, so nothing
+ is lost; the contract states that a phrase is the verb with the words that belong to it, so that
+ the statement reads on its own, and the judge reports "reads alone" beside "not stated"
+ ([extraction](extraction.md), #743).
+2. **A triple the sentence did not assert.** OLLIE showed how many extractions are conditioned,
+ attributed or hypothetical in the sentence and false without that context [Mausam 2012];
+ speculation is a property of the tuple, not of the sentence [Dong 2023b]. Qualifiers hold the
+ context [0037, 0044]. *Open:* a modality qualifier per statement (directive, plan, hypothesis),
+ and the rule that alignment never materializes a directive as a typed relation.
+3. **Minimal is not the same as self-contained.** MinIE says minimization loses context in its own
+ evaluation [Gashteovski 2017]; the neural systems went the other way and over-include
+ [Fatahi Bayat 2022]; CaRB [Bhardwaj 2019], WiRe57 [Léchelle 2019], LSOIE [Solawetz 2021] and
+ BenchIE [Gashteovski 2022] disagree on how compact a fact should be, and rankings flip with the
+ matcher. The fact-checking side reached the same place: fully atomic facts are not the right
+ unit, decontextuality and minimality pull against each other [Gunjal 2024, Choi 2021], and more
+ decomposition is not monotonically better [Hu 2025]. Our rule: one statement is one complete
+ proposition; coordinated objects split; a verb chain sharing one object does not.
+4. **The decomposition is itself a model output.** Change the decomposition prompt and every
+ downstream number moves [Wanner 2024]; a segmentation pre-pass can drop entities for strong
+ models [Pommeret 2026]. The contract is versioned in code and each bench round names its
+ contract [extraction](extraction.md). *Open:* entity recall measured with and without any
+ pre-pass before one is added.
+5. **Coordination and verb chains are the yield on dense sentences.** Splitting conjunctive
+ sentences first gave up to 1.8× yield with a precision gain [Saha 2018]; a coordination analyser
+ inside the extractor is what moved OpenIE6 [Kolluru 2020]; it can be done without a parser
+ [Wang 2023b]; split-and-rephrase is the general task [Narayan 2017]; irrelevant context is a
+ named failure of LLM OpenIE [Ling 2023]. The judge and the coverage script measure the dense
+ NVDA sentences separately (#731). *Open:* the figures dropped from dense sentences get a
+ second-look pass.
+6. **Common nouns as entities.** On Wikipedia text only 42% of open-extraction arguments are named
+ entities [Gashteovski 2019]; canonicalization meets the same string with different referents
+ first [Galárraga 2014]; the opposite failure, one referent under many strings, is the sparsity
+ KGGen was built against [Mo 2025], and coreference before extraction was the largest lever on
+ duplicate nodes in CORE-KG [Meher 2025]. "企业" merged across documents is a hub that means
+ nothing. A described thing exists only when a statement points at it [#736]. A thing is named
+ only by a proper name or a fixed term that means the same in any document; a role or a generic
+ phrase whose referent the passage decides is described, scoped to its document and never merged
+ by string; a generic phrase in an adjunct stays words in the qualifier (#743). The gate is a
+ term gate, not a proper-name gate: a disease, a drug or an indicator merges across documents by
+ design.
+7. **A closed schema makes the model invent.** Prompting with a fixed relation set causes
+ hallucination [Jiang 2024]; a model forced onto an ontology conforms almost perfectly and is
+ still wrong [van Cauter 2024]; the dominant "error" of LLM extraction against gold is spans the
+ gold never annotated [Han 2023]; exact match penalises paraphrase, so the honest evaluations are
+ human [Wadhwa 2023, Li 2023]. The open contract uses the document's words, the quote must occur
+ in the chunk, the run is at temperature 0, and the judge asks whether the text states it rather
+ than whether it matches a gold form; "not stated" stays under 2% on every corpus
+ ([extraction](extraction.md)). Hallucination and omission are reported as two numbers, as
+ Ghanem and Cruz ask [Ghanem 2024]: the judge and the coverage script.
+8. **A quote that is a substring is not yet support.** With frontier models 83 to 91% of required
+ quotes were even substrings and only 48 to 79% of those supported the claim [Windisch 2026];
+ citation presence is not citation support [Gao 2023]; a three-way label (attributable,
+ extrapolated, contradictory) is more useful than a binary one [Yue 2023]; abstention must be a
+ first-class output [Menick 2022]. `quote_not_in_chunk` is a drop reason and the judge is a
+ separate pass. *Open:* the judge's "misworded" splits into extrapolated and contradicted.
+9. **The contract's shape has a cost.** Format restrictions degrade reasoning [Tam 2024]; guideline
+ prompts are not reliably obeyed without training [Sainz 2024]; quoting before stating improves
+ grounding [Huang 2024]; forcing citations finer than a sentence hurts [Wang 2026]. The compact
+ contract was chosen by measurement (#731). *Open:* the quote comes before the statement in `s`;
+ clause-level quotes measured against minimal ones.
+10. **Numbers need their own patterns.** Open numerical extraction is its own problem
+ [Saha 2017]; the dense-sentence figures we drop are that class. *Open:* the second-look pass of
+ item 5.
+
+### Alignment
+
+11. **Arguments before relations.** Galárraga et al. cluster noun phrases first and relation
+ phrases on a semi-canonicalized base; four coarse argument types raised pairwise precision from
+ 0.946 to 0.997 because polysemous phrases split by signature [Galárraga 2014]; CESI inherits the
+ order [Vashishth 2018a]; the mapping of an open relation to a DBpedia property depends on the
+ argument types [Dutta 2015]; Angeli's KBP mapping is conditioned on the type signature
+ [Angeli 2015]; typed markers lift relation extraction [Ling 2012, Zhong 2021]; ODKE+ exposes only
+ the type's slice of the ontology [Khorshidi 2025]; iText2KG lists typing as its missing
+ ingredient [Lairgi 2024]. Kind words bind before relation phrases [0044 cut 2]. *Open:* the
+ phrase binding reads the classes at both ends, and the (class, class) signature is part of a
+ phrase binding's key.
+12. **Precision is lost in the binding, not in the extraction.** Of the mapping errors in the
+ three-hour KBP system, 31% were definition mismatch and 23% over-generalized rules against 15%
+ open-extraction errors [Soderland 2013]; the top hundred predicates cover only 57 to 82% of an
+ open base's triples [Romero 2023]; an open statement often maps to a conjunction of ontology
+ facts, not one property [Gashteovski 2020]. A word with no fitting class stays untyped and its
+ statements stay open; the only output is a suggestion to add a class ([ontology](ontology.md),
+ #741). *Open:* a phrase may bind to a formula, not only a property.
+13. **Nothing aligns to an undefined class.** EDC's gain comes from definitions written before
+ matching and a retriever whose recall bounds what can be bound [Zhang 2024]; definitions help
+ both retrieval and the judge [Song 2025]; a class is introduced by its description alone
+ [Obeidat 2019] and improved by example names [Ouyang 2024]; written guidelines are what make
+ zero-shot extraction work [Sainz 2024]. The signature the aligner reads is the kind word, its
+ spellings, example names and the relation phrases its things take part in; candidates come
+ from the ontology's embedding index when the base has one, else the whole list when it is small
+ (#741). *Open:* a class without a definition is not a binding target, and the ontology page
+ says so.
+14. **The characteristic error is the wrong level, not the wrong branch.** Naive LLM matching binds
+ a subclass to its parent [Norouzi 2023]; across ten models the errors are align-up, align-down
+ and disputed, with few outright false mappings [Qiang 2025]; order sensitivity concentrates
+ where the model is torn between its top options [Pezeshkpour 2024]; resolving top-down through
+ the hierarchy helps [Komarlu 2024]. *Open:* the queue card for an undecided word shows each
+ candidate with its parent and children; a candidate list includes the parent of every
+ candidate.
+15. **The judge prefers the first option.** Swapping two answers changed GPT-4's verdict in a third
+ of cases and Claude-v1's in three quarters [Zheng 2023]; reordering alone reversed a
+ leaderboard [Wang 2024]; selection bias comes from the prior on option-ID tokens [Zheng 2024].
+ Two votes, the second with candidates reversed; candidates are named by class key, never by a
+ letter; disagreement is a queue item, not a binding (#741). Spending the model's votes only on
+ borderline cases is the next economy [Taboada 2025].
+16. **Conformance is not correctness.** Ontology conformance near 1.0 with hallucinated arguments
+ [van Cauter 2024]; the standard metric names are conformance and hallucination rate
+ [Mihindukulasooriya 2023]; the standard check for the no-match case is a rejection rate over
+ concepts that have no counterpart [He 2023], now part of the OAEI Bio-ML evaluation [OAEI 2024].
+ *Open:* the alignment bench seeds kind words that fit no class and counts how often the aligner
+ binds them anyway.
+17. **Candidate recall is the ceiling.** Whatever the retriever misses can never be bound [Hertling
+ 2023, Zhang 2024]; the similarity threshold is a precision dial tuned per ontology [Qiang 2024];
+ the number of clusters is estimated, not set [Shen 2022]. The whole list goes to the model when
+ the base has at most sixty classes. *Open:* candidate recall measured on the batch base before
+ the embedding index is trusted above sixty.
+18. **A role is not a kind.** A class that names the role a thing plays in a relation cannot be
+ assigned from the thing's own kind word; Soderland et al. hit it with BombingAgent
+ [Soderland 2010]. *Open:* classes declared as roles are excluded from kind-word candidates.
+19. **A pipeline that cannot be re-run drifts.** Mappings are partial and revisable by design
+ [Madhavan 2007]; error propagation from extraction into fusion is the survey's standing warning
+ [Bian 2025]; the surveys ask for provenance and re-runnable stages [Weikum 2021, Hogan 2021].
+ Every binding carries its two votes and a `decided_at` compared with the class's `updated_at`;
+ only stale bindings are re-decided; every statement keeps its quote and offsets.
+
+### Time
+
+20. **The document type chooses the anchor, and the anchor is where the value goes wrong.** With
+ the same extractor, anchoring narrative text to the document's date instead of the previous
+ mention drops the value score from 89.8 to 64.9, and anchoring clinical-trial text to a local
+ time point zero raises it from 74.4 to 88.7 [Strötgen 2012]; in clinical narratives the anchor
+ choice is the least accurate component, at 74.7% [Sun 2015], and anchor errors are the top
+ cause of wrong values [Olex 2022]; a discharge summary has two legitimate document times,
+ admission and discharge [Sun 2013]. The document context is read once, mention anchors are
+ resolved in a second pass, and the anchor is stored in the interpretation, so a value always
+ says what it was computed from [0045]. *Open:* the strategy is chosen per document type; a
+ trial-relative mention ("week 4") is kept as an offset from a named local anchor in the manner
+ of HeidelTime's time point zero and THYME's event-anchored expressions [Styler 2014], so that it
+ becomes grade B the day the anchor is dated, instead of grade C forever.
+21. **Normalizing is harder than finding.** In TempEval-3 the best system found mentions at 90.3
+ but gave the right value at 77.6, and every participant normalized by rules [UzZaman 2013];
+ the neural parsers score operators such as "last" and "next" at a third of the rest
+ [Laparra 2018]; relative and underspecified mentions stay the hardest for prompted models
+ [Gautam 2024]. The model gives only an interpretation; code computes the value; nothing is
+ written when it cannot [0045]. The strongest recent version of that split generates a program
+ and executes it [Su 2025a]; the vocabulary of the code side is the ten operations of ARTime
+ [Ding 2021]. *Open:* the time bench scores a resolved interval by overlap with the reference,
+ not by string equality [Laparra 2018]; an ambiguous mention may return alternatives rather than
+ one value [Bethard 2013].
+22. **A wrong document date poisons every relative expression.** No study manipulates the document
+ date alone; the anchor numbers of item 20 are the quantification. The document's own stated
+ dates are the best evidence of its date [Chambers 2012]; year-level dating is useless for a
+ quarter [Vashishth 2018b] and dating from implicit cues is decades coarse [Kumar 2012]. The
+ agent-memory graphs of 2025 anchor relative dates to the message timestamp and collapse a date
+ to midnight [Rasmussen 2025], the two errors this ledger forbids; ATOM separates observed time
+ from valid time and measures run-to-run stability [Lairgi 2025]. Upload time and file
+ modification time are never a document date [0045, #714]; anything computed from the document
+ date is grade B. *Open:* a person changing the document date triggers re-resolution
+ (0045 cut 4); run-to-run stability of the interpretation call is a bench number.
+23. **A fiscal period is its own granularity.** Fiscal quarters and 52/53-week years are
+ non-regular granularities, and truncation to a calendar month is only defined when the target
+ groups the source [Bettini 2000]; a 52/53-week quarter truncated to the month is off by up to
+ six days. SEC filings carry the fiscal year end and a period code (FY, Q1 to Q4, H1, H2, M9, T1
+ to T3) in their XBRL metadata [SEC FSDS]. Fiscal quarters are computed from the fiscal year end
+ the opening states [0045]. *Open:* half-years and nine-month periods are period codes the
+ resolver accepts; the fiscal year end is read from the filing's metadata when the source is
+ EDGAR; fiscal periods are one granularity system beside the calendar, not a calendar month.
+24. **Mixing the two axes corrupts history.** Snodgrass's book is one long warning
+ [Snodgrass 1999]; transaction time is system-assigned and append-only, valid time is asserted
+ [Jensen 1999]; record time is the commit instant, assigned once and never back-dated
+ [Torp 2000]; SQL:2011 periods have neither a granularity nor an unknown end [Kulkarni 2012].
+ Two axes with separate predicates [0019]; the precision columns and the ended-unknown shape are
+ stated as extensions beyond the standard. *Open:* [time](time.md) names the axes once in the
+ glossary's words, valid and transaction [Jensen 1998].
+25. **Unknown is three different things.** Wikidata distinguishes a value, an unknown value that
+ exists, and no value; an end time is "unknown value" when the thing ended on an unknown date
+ and absent when it is ongoing [Wikidata Help]; the TimeBank anchoring corpus annotates a point,
+ an interval, a lower bound, an upper bound or nothing [Reimers 2016]; YAGO2 leaves a fact
+ without a time rather than give it an ill-defined one [Hoffart 2013]. Ours are `until`,
+ `ended_unknown` with `attested_to`, and an open end with `attested_to` [0022, 0045], and grade C
+ writes nothing. `attested_to` means true as of the last attestation, a now-relative value that
+ is stored as an explicit timestamp and evaluated when asked [Clifford 1997]; a fact whose last
+ attestation is old is stale, not false [Soulard 2025].
+26. **A table is a structure, not a paragraph.** The financial-report question-answering sets
+ feed a table to the model one row at a time with its row and column headers written out
+ [Chen 2021, Zhu 2021], and the pretrained table models flatten it without losing which row and
+ column a cell sits in [Herzig 2020]; the standard for turning cells into one faithful sentence
+ is ToTTo [Parikh 2020]; the statistical-report tables with hierarchical headers have their own
+ dataset, whose difficulty is "complex hierarchical indexing" [Cheng 2022, Zhao 2022]; how a
+ table is serialised, ordered and partitioned changes what a model reads from it [Sui 2024], a
+ change of structure with the same content costs accuracy until the structure is normalised
+ [Liu 2023], a whole table is not fed to a model at all when it is large [Chen 2024c], and the
+ survey of the field is [Fang 2024]; which column is the subject and which are properties is
+ the semantic table interpretation task [Cafarella 2008, SemTab]. Under #743 the earnings release lost its column headings between the HTML converter
+ and the chunker and the model wrote the period as the phrase. Tables are now rendered from the
+ DOM by structure, one row per line under its real headings, the section path folded into the
+ label, the caption run kept with every chunk ([sources](sources.md), #744); a heading that
+ names a time is `when`, a heading that names a change between periods stays the phrase; a
+ phrase equal to the value or to the subject is kept and counted. *Open:*
+ the DOCX, PDF and spreadsheet parsers render their tables the same way; the paragraph units of
+ the NVDA cross-domain corpus are recut from rendered rows with their headings, and the reference
+ facts carry the period; a check that a cell statement under a period column carries `when`.
+27. **Precision is not confidence, and vague is not coarse.** An indeterminate instant is a set of
+ chronons with a distribution, and comparing two of them needs a stated semantics
+ [Dyreson 1998]; confidence and bitemporality are separate axes [Chekol 2018]; "early 2019" is a
+ modifier with a calibrated width, not a coarser precision [Tissot 2019]; grading a date by the
+ provenance of its anchor is uncertainty typing, which the survey prefers to a scalar
+ [Jarnac 2025]. A precision column on every bound, with CHECK constraints that truncate the
+ value [0024]; the grade is a letter beside the value, never a probability. *Open:* the
+ cross-precision comparison used by the temporal engine, overlap or containment, is written
+ down in [time](time.md); a vague modifier is a qualifier on the mention, not a precision.
+
+## What has no precedent
+
+These decisions have no paper behind them and are held up only by our own benches
+([extraction](extraction.md), the batch review of #731, #740 and #741):
+
+- A described thing becomes an entity only when a statement points at it [#736].
+- The contract names things by their words, not by ids; ids mixed on dense passages [#731].
+- Temperature 0 and a verbatim quote as the condition of entry; "not stated" ≤ 2% on every corpus.
+- The three grades of a resolved time as a letter beside the value [0045]. YAGO2's refusal to write
+ an ill-defined time and the survey's preference for typed uncertainty are precedents in spirit,
+ not in form.
+- Every intermediate layer stays in the ledger and is re-run by a job, instead of a pipeline that
+ computes once.
+
+Three things the survey looked for and did not find: a controlled comparison of schema-first and
+schema-late extraction with the same extractor; a paper that studies generic-noun hubs as a named
+phenomenon; a study that manipulates the document date alone and measures what it does to relative
+mentions. All three are ours to measure.
+
+## References
+
+Every entry was verified against the linked page. Grouped by the layer that cites it; a paper cited
+by two layers is listed once.
+
+### Open extraction
+
+- [Banko 2007] Banko, Cafarella, Soderland, Broadhead, Etzioni. Open Information Extraction from the Web. IJCAI 2007. https://www.ijcai.org/Proceedings/07/Papers/429.pdf
+- [Fader 2011] Fader, Soderland, Etzioni. Identifying Relations for Open Information Extraction. EMNLP 2011. https://aclanthology.org/D11-1142/
+- [Mausam 2012] Mausam, Schmitz, Bart, Soderland, Etzioni. Open Language Learning for Information Extraction. EMNLP-CoNLL 2012. https://aclanthology.org/D12-1048/
+- [Del Corro 2013] Del Corro, Gemulla. ClausIE: Clause-Based Open Information Extraction. WWW 2013. https://dl.acm.org/doi/10.1145/2488388.2488420
+- [Angeli 2015] Angeli, Johnson Premkumar, Manning. Leveraging Linguistic Structure For Open Domain Information Extraction. ACL 2015. https://aclanthology.org/P15-1034/
+- [Stanovsky 2018] Stanovsky, Michael, Zettlemoyer, Dagan. Supervised Open Information Extraction. NAACL 2018. https://aclanthology.org/N18-1081/
+- [Kolluru 2020] Kolluru, Adlakha, Aggarwal, Mausam, Chakrabarti. OpenIE6: Iterative Grid Labeling and Coordination Analysis for Open Information Extraction. EMNLP 2020. https://aclanthology.org/2020.emnlp-main.306/
+- [Niklaus 2018] Niklaus, Cetto, Freitas, Handschuh. A Survey on Open Information Extraction. COLING 2018. https://aclanthology.org/C18-1326/
+- [Zhou 2022] Zhou et al. A Survey on Neural Open Information Extraction: Current Status and Future Directions. IJCAI 2022. https://www.ijcai.org/proceedings/2022/793
+- [Liu 2024] Liu et al. A Survey on Open Information Extraction from Rule-based Model to Large Language Model. Findings of EMNLP 2024. https://aclanthology.org/2024.findings-emnlp.560/
+- [Pei 2023] Pei et al. Abstractive Open Information Extraction. EMNLP 2023. https://aclanthology.org/2023.emnlp-main.376/
+- [Tseng 2014] Tseng, Lee, Lin, Liao, Liu, Chen, Etzioni, Fader. Chinese Open Relation Extraction for Knowledge Acquisition. EACL 2014. https://aclanthology.org/E14-4003/
+- [Qiu 2014] Qiu, Zhang. ZORE: A Syntax-based System for Chinese Open Relation Extraction. EMNLP 2014. https://aclanthology.org/D14-1201/
+- [Sun 2018] Sun, Li, Wang, Fan, Feng, Li. Logician: A Unified End-to-End Neural Approach for Open-Domain Information Extraction. WSDM 2018. https://arxiv.org/abs/1904.12535
+- [Deng 2022] Deng et al. Title2Event: Benchmarking Open Event Extraction with a Large-scale Chinese Title Dataset. EMNLP 2022. https://aclanthology.org/2022.emnlp-main.437/
+- [Abolhasani 2026] Abolhasani, Ba, He, Pan. Beyond Predefined Schemas: TRACE-KG for Context-Enriched Knowledge Graph Generation. 2026. https://arxiv.org/abs/2604.03496
+- [Han 2026] Han, Lai. From Symbolic to Natural-Language Relations: Rethinking Knowledge Graph Construction in the Era of Large Language Models. 2026. https://arxiv.org/abs/2601.09069
+- [Dong 2023a] Dong et al. Open Information Extraction via Chunks. EMNLP 2023. https://aclanthology.org/2023.emnlp-main.951/
+- [Dong 2023b] Dong et al. From Speculation Detection to Trustworthy Relational Tuples in Information Extraction. Findings of EMNLP 2023. https://aclanthology.org/2023.findings-emnlp.886/
+- [Ling 2023] Ling et al. Improving Open Information Extraction with Large Language Models: A Study on Demonstration Uncertainty. 2023. https://arxiv.org/abs/2309.03433
+- [Saha 2017] Saha, Pal, Mausam. Bootstrapping for Numerical Open IE. ACL 2017. https://aclanthology.org/P17-2050/
+- [Saha 2018] Saha, Mausam. Open Information Extraction from Conjunctive Sentences. COLING 2018. https://aclanthology.org/C18-1194/
+- [Wang 2023b] Wang et al. CoRec: An Easy Approach for Coordination Recognition. EMNLP 2023. https://aclanthology.org/2023.emnlp-main.934/
+- [Narayan 2017] Narayan, Gardent, Cohen, Shimorina. Split and Rephrase. EMNLP 2017. https://aclanthology.org/D17-1064/
+
+### Granularity, qualifiers, propositions
+
+- [Gashteovski 2017] Gashteovski, Gemulla, Del Corro. MinIE: Minimizing Facts in Open Information Extraction. EMNLP 2017. https://aclanthology.org/D17-1278/
+- [Bhutani 2016] Bhutani, Jagadish, Radev. Nested Propositions in Open Information Extraction. EMNLP 2016. https://aclanthology.org/D16-1006/
+- [Niklaus 2019] Niklaus, Cetto, Freitas, Handschuh. Transforming Complex Sentences into a Semantic Hierarchy. ACL 2019. https://aclanthology.org/P19-1333/
+- [Vrandečić 2014] Vrandečić, Krötzsch. Wikidata: A Free Collaborative Knowledgebase. Communications of the ACM 57(10), 2014. https://dl.acm.org/doi/10.1145/2629489
+- [Chia 2022] Chia et al. A Dataset for Hyper-Relational Extraction and a Cube-Filling Approach. EMNLP 2022. https://aclanthology.org/2022.emnlp-main.688/
+- [Luo 2024] Luo et al. Text2NKG: Fine-Grained N-ary Relation Extraction for N-ary relational Knowledge Graph Construction. NeurIPS 2024. https://arxiv.org/abs/2310.05185
+- [Fatahi Bayat 2022] Fatahi Bayat, Bhutani, Jagadish. CompactIE: Compact Facts in Open Information Extraction. NAACL 2022. https://aclanthology.org/2022.naacl-main.65/
+- [Bhardwaj 2019] Bhardwaj, Aggarwal, Mausam. CaRB: A Crowdsourced Benchmark for Open IE. EMNLP-IJCNLP 2019. https://aclanthology.org/D19-1651/
+- [Léchelle 2019] Léchelle, Gotti, Langlais. WiRe57: A Fine-Grained Benchmark for Open Information Extraction. LAW 2019. https://aclanthology.org/W19-4002/
+- [Solawetz 2021] Solawetz, Larson. LSOIE: A Large-Scale Dataset for Supervised Open Information Extraction. EACL 2021. https://aclanthology.org/2021.eacl-main.222/
+- [Gashteovski 2022] Gashteovski, Yu, Kotnis, Lawrence, Niepert, Glavaš. BenchIE: A Framework for Multi-Faceted Fact-Based Open Information Extraction Evaluation. ACL 2022. https://aclanthology.org/2022.acl-long.307/
+- [Lamarche 2024] Lamarche, Langlais. BenchIE^FL: A Manually Re-Annotated Fact-Based Open Information Extraction Benchmark. Findings of ACL 2024. https://aclanthology.org/2024.findings-acl.496/
+- [Gunjal 2024] Gunjal, Durrett. Molecular Facts: Desiderata for Decontextualization in LLM Fact Verification. Findings of EMNLP 2024. https://aclanthology.org/2024.findings-emnlp.215/
+- [Choi 2021] Choi, Palomaki, Lamm, Kwiatkowski, Das, Collins. Decontextualization: Making Sentences Stand-Alone. TACL 2021. https://aclanthology.org/2021.tacl-1.27/
+- [Hu 2025] Hu et al. Decomposition Dilemmas: Does Claim Decomposition Boost or Burden Fact-Checking Performance? NAACL 2025. https://aclanthology.org/2025.naacl-long.320/
+- [Wanner 2024] Wanner et al. A Closer Look at Claim Decomposition. *SEM 2024. https://aclanthology.org/2024.starsem-1.13/
+- [Pommeret 2026] Pommeret et al. LLM-based Atomic Propositions help weak extractors: Evaluation of a Propositioner for triplet extraction. 2026. https://arxiv.org/abs/2604.02866
+- [Min 2023] Min et al. FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation. EMNLP 2023. https://aclanthology.org/2023.emnlp-main.741/
+- [Chen 2024a] Chen et al. Dense X Retrieval: What Retrieval Granularity Should We Use? EMNLP 2024. https://aclanthology.org/2024.emnlp-main.845/
+
+### Grounding and evaluation of model extraction
+
+- [Jiang 2024] Jiang et al. GenRES: Rethinking Evaluation for Generative Relation Extraction in the Era of Large Language Models. NAACL 2024. https://aclanthology.org/2024.naacl-long.155/
+- [Han 2023] Han et al. Is Information Extraction Solved by ChatGPT? An Analysis of Performance, Evaluation Criteria, Robustness and Errors. 2023. https://arxiv.org/abs/2305.14450
+- [Wadhwa 2023] Wadhwa, Amir, Wallace. Revisiting Relation Extraction in the era of Large Language Models. ACL 2023. https://aclanthology.org/2023.acl-long.868/
+- [Li 2023] Li et al. Evaluating ChatGPT's Information Extraction Capabilities: An Assessment of Performance, Explainability, Calibration, and Faithfulness. 2023. https://arxiv.org/abs/2304.11633
+- [Ghanem 2024] Ghanem, Cruz. Enhancing Knowledge Graph Construction: Evaluating with Emphasis on Hallucination, Omission, and Graph Similarity Metrics. KGSWC 2024. https://arxiv.org/abs/2502.05239
+- [Windisch 2026] Windisch et al. Show Your Work: Verbatim Evidence Requirements and Automated Assessment of Large Language Models for Biomedical Text Processing of Trial Eligibility Criteria. Cureus 2026. https://pmc.ncbi.nlm.nih.gov/articles/PMC13252682/
+- [Gao 2023] Gao, Yen, Yu, Chen. Enabling Large Language Models to Generate Text with Citations. EMNLP 2023. https://aclanthology.org/2023.emnlp-main.398/
+- [Yue 2023] Yue et al. Automatic Evaluation of Attribution by Large Language Models. Findings of EMNLP 2023. https://aclanthology.org/2023.findings-emnlp.307/
+- [Menick 2022] Menick et al. Teaching language models to support answers with verified quotes. 2022. https://arxiv.org/abs/2203.11147
+- [Huang 2024] Huang et al. Learning Fine-Grained Grounded Citations for Attributed Large Language Models. Findings of ACL 2024. https://aclanthology.org/2024.findings-acl.838/
+- [Zhang 2025] Zhang et al. Verifiable by Design: Aligning Language Models to Quote from Pre-Training Data. NAACL 2025. https://aclanthology.org/2025.naacl-long.191/
+- [Tam 2024] Tam et al. Let Me Speak Freely? A Study on the Impact of Format Restrictions on Large Language Model Performance. EMNLP 2024 Industry Track. https://aclanthology.org/2024.emnlp-industry.91/
+- [Sainz 2024] Sainz et al. GoLLIE: Annotation Guidelines improve Zero-Shot Information-Extraction. ICLR 2024. https://arxiv.org/abs/2310.03668
+- [Wang 2026] Wang et al. Are Finer Citations Always Better? Rethinking Granularity for Attributed Generation. 2026. https://arxiv.org/abs/2604.01432
+
+### Canonicalization and alignment
+
+- [Galárraga 2014] Galárraga, Heitz, Murphy, Suchanek. Canonicalizing Open Knowledge Bases. CIKM 2014. https://dl.acm.org/doi/10.1145/2661829.2662073
+- [Vashishth 2018a] Vashishth, Jain, Talukdar. CESI: Canonicalizing Open Knowledge Bases using Embeddings and Side Information. WWW 2018. https://dl.acm.org/doi/10.1145/3178876.3186030
+- [Dash 2021] Dash, Rossiello, Mihindukulasooriya, Bagchi, Gliozzo. Open Knowledge Graphs Canonicalization using Variational Autoencoders. EMNLP 2021. https://aclanthology.org/2021.emnlp-main.811/
+- [Liu 2021] Liu et al. Joint Open Knowledge Base Canonicalization and Linking. SIGMOD 2021. https://dl.acm.org/doi/10.1145/3448016.3452776
+- [Shen 2022] Shen, Yang, Liu. Multi-View Clustering for Open Knowledge Base Canonicalization. KDD 2022. https://dl.acm.org/doi/10.1145/3534678.3539449
+- [Yang 2024] Yang, Curry. Open Knowledge Base Canonicalization: Techniques and Challenges. TEXT2KG 2024 (ESWC), CEUR Vol-3747. https://ceur-ws.org/Vol-3747/text2kg_paper5.pdf
+- [Gashteovski 2019] Gashteovski, Wanner, Hertling, Broscheit, Gemulla. OPIEC: An Open Information Extraction Corpus. AKBC 2019. https://arxiv.org/abs/1904.12324
+- [Gashteovski 2020] Gashteovski et al. On Aligning OpenIE Extractions with Knowledge Bases: A Case Study. Eval4NLP 2020. https://aclanthology.org/2020.eval4nlp-1.14/
+- [Soderland 2010] Soderland, Roof, Qin, Xu, Mausam, Etzioni. Adapting Open Information Extraction to Domain-Specific Relations. AI Magazine 31(3), 2010. https://ojs.aaai.org/index.php/aimagazine/article/view/2305
+- [Soderland 2013] Soderland, Gilmer, Bart, Etzioni, Weld. Open Information Extraction to KBP Relations in 3 Hours. TAC 2013. https://tac.nist.gov/publications/2013/participant.papers/UWashington.TAC2013.proceedings.pdf
+- [Dutta 2015] Dutta, Meilicke, Stuckenschmidt. Enriching Structured Knowledge with Open Information. WWW 2015. https://dl.acm.org/doi/10.1145/2736277.2741139
+- [Romero 2023] Romero, Razniewski. Mapping and Cleaning Open Commonsense Knowledge Bases with Generative Translation. ISWC 2023. https://arxiv.org/abs/2306.12766
+- [Zhang 2024] Zhang, Soh. Extract, Define, Canonicalize: An LLM-based Framework for Knowledge Graph Construction. EMNLP 2024. https://aclanthology.org/2024.emnlp-main.548/
+- [Mo 2025] Mo et al. KGGen: Extracting Knowledge Graphs from Plain Text with Language Models. NeurIPS 2025. https://arxiv.org/abs/2502.09956
+- [Lairgi 2024] Lairgi, Moncla, Cazabet, Benabdeslem, Cléau. iText2KG: Incremental Knowledge Graphs Construction Using Large Language Models. WISE 2024. https://arxiv.org/abs/2409.03284
+- [Meher 2025] Meher, Domeniconi. Inside CORE-KG: Evaluating Structured Prompting and Coreference Resolution for Knowledge Graphs. 2025. https://arxiv.org/abs/2510.26512
+- [Bai 2025] Bai et al. AutoSchemaKG: Autonomous Knowledge Graph Construction through Dynamic Schema Induction from Web-Scale Corpora. 2025. https://arxiv.org/abs/2505.23628
+- [Chen 2024b] Chen et al. SAC-KG: Exploiting Large Language Models as Skilled Automatic Constructors for Domain Knowledge Graph. ACL 2024. https://aclanthology.org/2024.acl-long.238/
+- [Sun 2025] Sun et al. Docs2KG: A Human-LLM Collaborative Approach to Unified Knowledge Graph Construction from Heterogeneous Documents. WWW Companion 2025. https://arxiv.org/abs/2406.02962
+- [Khorshidi 2025] Khorshidi et al. ODKE+: Ontology-Guided Open-Domain Knowledge Extraction with LLMs. 2025. https://arxiv.org/abs/2509.04696
+- [Bian 2025] Bian. LLM-empowered knowledge graph construction: A survey. 2025. https://arxiv.org/abs/2510.20345
+- [Mihindukulasooriya 2023] Mihindukulasooriya, Tiwari, Enguix, Lata. Text2KGBench: A Benchmark for Ontology-Driven Knowledge Graph Generation from Text. ISWC 2023. https://arxiv.org/abs/2308.02357
+- [van Cauter 2024] van Cauter, Yakovets. Ontology-guided Knowledge Graph Construction from Maintenance Short Texts. KaLLM 2024 (ACL). https://aclanthology.org/2024.kallm-1.8/
+
+### Typing
+
+- [Choi 2018] Choi, Levy, Choi, Zettlemoyer. Ultra-Fine Entity Typing. ACL 2018. https://aclanthology.org/P18-1009/
+- [Dai 2023] Dai, Zeng. From Ultra-Fine to Fine: Fine-tuning Ultra-Fine Entity Typing Models to Fine-grained. ACL 2023. https://aclanthology.org/2023.acl-long.126/
+- [Obeidat 2019] Obeidat, Fern, Shahbazi, Tadepalli. Description-Based Zero-shot Fine-Grained Entity Typing. NAACL 2019. https://aclanthology.org/N19-1087/
+- [Ouyang 2024] Ouyang, Huang, Pillai, Zhang, Zhang, Han. Ontology Enrichment for Effective Fine-grained Entity Typing. KDD 2024. https://dl.acm.org/doi/10.1145/3637528.3671857
+- [Komarlu 2024] Komarlu, Jiang, Wang, Han. OntoType: Ontology-Guided and Pre-Trained Language Model Assisted Fine-Grained Entity Typing. KDD 2024. https://dl.acm.org/doi/10.1145/3637528.3671745
+- [Ling 2012] Ling, Weld. Fine-Grained Entity Recognition. AAAI 2012. https://ojs.aaai.org/index.php/AAAI/article/view/8122
+- [Zhong 2021] Zhong, Chen. A Frustratingly Easy Approach for Entity and Relation Extraction. NAACL 2021. https://aclanthology.org/2021.naacl-main.5/
+
+### Matching and judging
+
+- [He 2023] He, Chen, Dong, Horrocks. Exploring Large Language Models for Ontology Alignment. ISWC 2023 Posters & Demos. https://arxiv.org/abs/2309.07172
+- [Hertling 2023] Hertling, Paulheim. OLaLa: Ontology Matching with Large Language Models. K-CAP 2023. https://dl.acm.org/doi/10.1145/3587259.3627571
+- [Babaei Giglou 2024] Babaei Giglou, D'Souza, Engel, Auer. LLMs4OM: Matching Ontologies with Large Language Models. 2024. https://arxiv.org/abs/2404.10317
+- [Qiang 2024] Qiang, Wang, Taylor. Agent-OM: Leveraging LLM Agents for Ontology Matching. PVLDB 18(3), 2024. https://www.vldb.org/pvldb/vol18/p516-qiang.pdf
+- [Norouzi 2023] Norouzi, Mahdavinejad, Hitzler. Conversational Ontology Alignment with ChatGPT. 2023. https://arxiv.org/abs/2308.09217
+- [Qiang 2025] Qiang, Taylor, Wang, Jiang. OAEI-LLM-T: A TBox Benchmark Dataset for Understanding Large Language Model Hallucinations in Ontology Matching. 2025. https://arxiv.org/abs/2503.21813
+- [OAEI 2024] Abd Nikooie Pour et al. Results of the Ontology Alignment Evaluation Initiative 2024. OM 2024 (ISWC), CEUR Vol-3897. https://ceur-ws.org/Vol-3897/oaei2024_paper0.pdf
+- [Taboada 2025] Taboada, Martinez, Arideh, Mosquera. Ontology Matching with Large Language Models and Prioritized Depth-First Search. 2025. https://arxiv.org/abs/2501.11441
+- [Song 2025] Song, Chen, Schmidt. GenOM: Ontology Matching with Description Generation and Large Language Model. OM 2025 (ISWC), CEUR Vol-4144. https://ceur-ws.org/Vol-4144/om2025-LTpaper3.pdf
+- [Zheng 2023] Zheng et al. Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena. NeurIPS 2023 Datasets and Benchmarks. https://arxiv.org/abs/2306.05685
+- [Wang 2024] Wang et al. Large Language Models are not Fair Evaluators. ACL 2024. https://aclanthology.org/2024.acl-long.511/
+- [Pezeshkpour 2024] Pezeshkpour, Hruschka. Large Language Models Sensitivity to The Order of Options in Multiple-Choice Questions. Findings of NAACL 2024. https://aclanthology.org/2024.findings-naacl.130/
+- [Zheng 2024] Zheng, Zhou, Meng, Zhou, Huang. Large Language Models Are Not Robust Multiple Choice Selectors. ICLR 2024. https://arxiv.org/abs/2309.03882
+- [Wang 2023a] Wang et al. Self-Consistency Improves Chain of Thought Reasoning in Language Models. ICLR 2023. https://arxiv.org/abs/2203.11171
+
+### Integration and surveys
+
+- [Franklin 2005] Franklin, Halevy, Maier. From Databases to Dataspaces: A New Abstraction for Information Management. SIGMOD Record 34(4), 2005. https://dl.acm.org/doi/10.1145/1107499.1107502
+- [Halevy 2006] Halevy, Franklin, Maier. Principles of Dataspace Systems. PODS 2006. https://dl.acm.org/doi/10.1145/1142351.1142352
+- [Madhavan 2007] Madhavan et al. Web-scale Data Integration: You can only afford to Pay As You Go. CIDR 2007. https://www.cidrdb.org/cidr2007/papers/cidr07p40.pdf
+- [Poggi 2008] Poggi, Lembo, Calvanese, De Giacomo, Lenzerini, Rosati. Linking Data to Ontologies. Journal on Data Semantics X, LNCS 4900, 2008. https://link.springer.com/chapter/10.1007/978-3-540-77688-8_5
+- [Xiao 2018] Xiao et al. Ontology-Based Data Access: A Survey. IJCAI 2018. https://www.ijcai.org/proceedings/2018/777
+- [Hogan 2021] Hogan et al. Knowledge Graphs. ACM Computing Surveys 54(4), 2021. https://arxiv.org/abs/2003.02320
+- [Weikum 2021] Weikum, Dong, Razniewski, Suchanek. Machine Knowledge: Creation and Curation of Comprehensive Knowledge Bases. Foundations and Trends in Databases 10(2–4), 2021. https://arxiv.org/abs/2009.11564
+
+### Tables
+
+- [Chen 2021] Chen et al. FinQA: A Dataset of Numerical Reasoning over Financial Data. EMNLP 2021. https://aclanthology.org/2021.emnlp-main.300/
+- [Zhu 2021] Zhu et al. TAT-QA: A Question Answering Benchmark on a Hybrid of Tabular and Textual Content in Finance. ACL-IJCNLP 2021. https://aclanthology.org/2021.acl-long.254/
+- [Parikh 2020] Parikh et al. ToTTo: A Controlled Table-To-Text Generation Dataset. EMNLP 2020. https://aclanthology.org/2020.emnlp-main.89/
+- [Herzig 2020] Herzig et al. TaPas: Weakly Supervised Table Parsing via Pre-training. ACL 2020. https://aclanthology.org/2020.acl-main.398/
+- [Cheng 2022] Cheng et al. HiTab: A Hierarchical Table Dataset for Question Answering and Natural Language Generation. ACL 2022. https://aclanthology.org/2022.acl-long.78/
+- [Zhao 2022] Zhao et al. MultiHiertt: Numerical Reasoning over Multi Hierarchical Tabular and Textual Data. ACL 2022. https://aclanthology.org/2022.acl-long.454/
+- [Cafarella 2008] Cafarella, Halevy, Wang, Wu, Zhang. WebTables: Exploring the Power of Tables on the Web. VLDB 2008. https://www.vldb.org/pvldb/vol1/1453916.pdf
+- [SemTab] Semantic Web Challenge on Tabular Data to Knowledge Graph Matching, ISWC, since 2019. https://www.cs.ox.ac.uk/isg/challenges/sem-tab/
+- [Sui 2024] Sui et al. Table Meets LLM: Can Large Language Models Understand Structured Table Data? A Benchmark and Empirical Study. WSDM 2024. https://arxiv.org/abs/2305.13062
+- [Liu 2023] Liu et al. Rethinking Tabular Data Understanding with Large Language Models. 2023. https://arxiv.org/abs/2312.16702
+- [Chen 2024c] Chen et al. TableRAG: Million-Token Table Understanding with Language Models. NeurIPS 2024. https://arxiv.org/abs/2410.04739
+- [Fang 2024] Fang et al. Large Language Models (LLMs) on Tabular Data: Prediction, Generation, and Understanding. A Survey. TMLR 2024. https://arxiv.org/abs/2402.17944
+
+### Time: tagging and normalization
+
+- [Pustejovsky 2003] Pustejovsky et al. TimeML: Robust Specification of Event and Temporal Expressions in Text. New Directions in Question Answering (AAAI Spring Symposium), 2003. https://aaai.org/papers/0005-ss03-07-005-timeml-robust-specification-of-event-and-temporal-expressions-in-text/
+- [UzZaman 2013] UzZaman, Llorens, Derczynski, Allen, Verhagen, Pustejovsky. SemEval-2013 Task 1: TempEval-3: Evaluating Time Expressions, Events, and Temporal Relations. SemEval 2013. https://aclanthology.org/S13-2001/
+- [Chang 2012] Chang, Manning. SUTime: A library for recognizing and normalizing time expressions. LREC 2012. https://aclanthology.org/L12-1122/
+- [Strötgen 2012] Strötgen, Gertz. Temporal Tagging on Different Domains: Challenges, Strategies, and Gold Standards. LREC 2012. https://aclanthology.org/L12-1219/
+- [Strötgen 2013] Strötgen, Gertz. Multilingual and cross-domain temporal tagging. Language Resources and Evaluation 47(2), 2013. https://link.springer.com/article/10.1007/s10579-012-9179-y
+- [Bethard 2013] Bethard. A Synchronous Context Free Grammar for Time Normalization. EMNLP 2013. https://aclanthology.org/D13-1078/
+- [Bethard 2016a] Bethard, Parker. A Semantically Compositional Annotation Scheme for Time Normalization. LREC 2016. https://aclanthology.org/L16-1599/
+- [Laparra 2018] Laparra, Xu, Bethard. From Characters to Time Intervals: New Paradigms for Evaluation and Neural Parsing of Time Normalizations. TACL 6, 2018. https://aclanthology.org/Q18-1025/
+- [Ding 2021] Ding, Chen, Li, Qu. Automatic Rule Generation for Time Expression Normalization. Findings of EMNLP 2021. https://aclanthology.org/2021.findings-emnlp.269/
+- [Gautam 2024] Gautam, Lange, Strötgen. Discourse-Aware In-Context Learning for Temporal Expression Normalization. NAACL 2024. https://aclanthology.org/2024.naacl-short.27/
+- [Su 2025a] Su, Yu, Howard, Bethard. A Semantic Parsing Framework for End-to-End Time Normalization. NeurIPS 2025. https://arxiv.org/abs/2507.06450
+- [Su 2025b] Su, Howard, Bethard. Transformer-Based Temporal Information Extraction and Application: A Review. EMNLP 2025. https://aclanthology.org/2025.emnlp-main.1467/
+- [Zhong 2023] Zhong, Cambria. Time expression recognition and normalization: a survey. Artificial Intelligence Review 56(9), 2023. https://link.springer.com/article/10.1007/s10462-023-10400-y
+- [Leeuwenberg 2019] Leeuwenberg, Moens. A Survey on Temporal Reasoning for Temporal Information Extraction from Text. JAIR, 2019. https://jair.org/index.php/jair/article/view/11727
+
+### Time: document dates, validity, clinical relative time
+
+- [Chambers 2012] Chambers. Labeling Documents with Timestamps: Learning from their Time Expressions. ACL 2012. https://aclanthology.org/P12-1011/
+- [Vashishth 2018b] Vashishth, Dasgupta, Ray, Talukdar. Dating Documents using Graph Convolution Networks. ACL 2018. https://aclanthology.org/P18-1149/
+- [Kumar 2012] Kumar, Baldridge, Lease, Ghosh. Dating Texts without Explicit Temporal Cues. 2012. https://arxiv.org/abs/1211.2290
+- [Hoffart 2013] Hoffart, Suchanek, Berberich, Weikum. YAGO2: A spatially and temporally enhanced knowledge base from Wikipedia. Artificial Intelligence 194, 2013. https://doi.org/10.1016/j.artint.2012.06.001
+- [Ling 2010] Ling, Weld. Temporal Information Extraction. AAAI 2010. https://ojs.aaai.org/index.php/AAAI/article/view/7512
+- [Wang 2011] Wang, Yang, Qu, Spaniol, Weikum. Harvesting Facts from Textual Web Sources by Constrained Label Propagation. CIKM 2011. https://dl.acm.org/doi/10.1145/2063576.2063698
+- [Talukdar 2012] Talukdar, Wijaya, Mitchell. Coupled Temporal Scoping of Relational Facts. WSDM 2012. https://dl.acm.org/doi/10.1145/2124295.2124307
+- [Reimers 2016] Reimers, Dehghani, Gurevych. Temporal Anchoring of Events for the TimeBank Corpus. ACL 2016. https://aclanthology.org/P16-1207/
+- [Rasmussen 2025] Rasmussen, Paliychuk, Beauvais, Ryan, Chalef. Zep: A Temporal Knowledge Graph Architecture for Agent Memory. 2025. https://arxiv.org/abs/2501.13956
+- [Lairgi 2025] Lairgi, Moncla, Benabdeslem, Cazabet, Cléau. ATOM: AdapTive and OptiMized dynamic temporal knowledge graph construction using LLMs. 2025. https://arxiv.org/abs/2510.22590
+- [Soulard 2025] Soulard, Saïs, Raad. When Facts Expire: Benchmarking Temporal Validity in Knowledge Graphs. CIKM 2025. https://dl.acm.org/doi/10.1145/3746252.3761648
+- [Sun 2013] Sun, Rumshisky, Uzuner. Evaluating temporal relations in clinical text: 2012 i2b2 Challenge. JAMIA 20(5), 2013. https://academic.oup.com/jamia/article/20/5/806/726374
+- [Sun 2015] Sun, Rumshisky, Uzuner. Normalization of relative and incomplete temporal expressions in clinical narratives. JAMIA 22(5), 2015. https://academic.oup.com/jamia/article-abstract/22/5/1001/928475
+- [Styler 2014] Styler et al. Temporal Annotation in the Clinical Domain. TACL 2, 2014. https://aclanthology.org/Q14-1012/
+- [Olex 2022] Olex, McInnes. Temporal disambiguation of relative temporal expressions in clinical texts. Frontiers in Research Metrics and Analytics 7, 2022. https://www.frontiersin.org/journals/research-metrics-and-analytics/articles/10.3389/frma.2022.1001266/full
+- [SEC FSDS] U.S. Securities and Exchange Commission. Financial Statement Data Sets, readme (fields `fye`, `fy`, `fp`, `period`). https://www.sec.gov/files/aqfs.pdf
+- [Wikidata Help] Wikidata. Help:Statements (value, unknown value, no value); Help:Dates (precision, earliest and latest date); Property:P582 end time. https://www.wikidata.org/wiki/Help:Dates
+
+### Time: the ledger
+
+- [Snodgrass 1999] Snodgrass. Developing Time-Oriented Database Applications in SQL. Morgan Kaufmann, 1999. https://dl.acm.org/doi/10.5555/320037
+- [Jensen 1999] Jensen, Snodgrass. Temporal Data Management. IEEE TKDE 11(1), 1999. https://www2.cs.arizona.edu/~rts/pubs/TKDEJan99.pdf
+- [Jensen 1998] Jensen, Dyreson et al. The Consensus Glossary of Temporal Database Concepts, February 1998 Version. Temporal Databases: Research and Practice, LNCS 1399, 1998. https://link.springer.com/chapter/10.1007/BFb0053710
+- [Kulkarni 2012] Kulkarni, Michels. Temporal features in SQL:2011. SIGMOD Record 41(3), 2012. https://dl.acm.org/doi/10.1145/2380776.2380786
+- [Clifford 1997] Clifford, Dyreson, Isakowitz, Jensen, Snodgrass. On the Semantics of "Now" in Databases. ACM TODS 22(2), 1997. https://ulp.cs.arizona.edu/people/rts/pubs/TODS97.pdf
+- [Torp 2000] Torp, Jensen, Snodgrass. Effective Timestamping in Databases. VLDB Journal 8, 2000. https://link.springer.com/article/10.1007/s007780050008
+- [Bettini 2000] Bettini, Jajodia, Wang. Time Granularities in Databases, Data Mining, and Temporal Reasoning. Springer, 2000. https://link.springer.com/book/10.1007/978-3-662-04228-1
+- [Dyreson 1998] Dyreson, Snodgrass. Supporting Valid-time Indeterminacy. ACM TODS 23(1), 1998. https://dl.acm.org/doi/10.1145/288086.288087
+- [Chekol 2018] Chekol, Stuckenschmidt. Towards Probabilistic Bitemporal Knowledge Graphs. WWW 2018 Companion. https://dl.acm.org/doi/10.1145/3184558.3191637
+- [Tissot 2019] Tissot, Del Fabro, Derczynski, Roberts. Normalisation of imprecise temporal expressions extracted from text. Knowledge and Information Systems 61(3), 2019. https://link.springer.com/article/10.1007/s10115-019-01338-1
+- [Jarnac 2025] Jarnac, Chabot, Couceiro. Uncertainty Management in the Construction of Knowledge Graphs: A Survey. TGDK 3(1), 2025. https://drops.dagstuhl.de/entities/document/10.4230/TGDK.3.1.3
diff --git a/docs/design/sources.md b/docs/design/sources.md
index 76aed25e..8a118a73 100644
--- a/docs/design/sources.md
+++ b/docs/design/sources.md
@@ -35,6 +35,22 @@ every table that follows it; a row is never split; headings become a breadcrumb
overlap; chunk text is verbatim slices so a quote is found by offset; a document is re-chunked on
its next reprocess [0039].
+**Tables** [#744, prior-work item 26]. An HTML table is rendered from the DOM before the Markdown
+converter sees it, by structure alone: cells hidden by style are skipped, spanning cells are laid on
+a grid, columns empty in every row go, a cell holding only a symbol ("$", ")") rejoins the number
+beside it. Rows are classified by what they hold: a lone cell before any header or data is a
+caption line, a row of words with no label column is a header row, a lone label after the headers
+is a section, a label followed by figures is data. Header rows are stacked per column into one
+heading ("Three Months Ended July 26, 2026"); sections fold into the labels of the rows under them
+("Current assets › Accounts receivable, net"), with indentation read from leading empty cells or
+the label's padding, and a section ends at a row of its own depth only once it has had a deeper
+child; caption lines become a paragraph before the table that ends with a colon. The chunker then
+does what it already did: the caption and the header travel with every piece, and a run of up to
+five short paragraphs before a table is its caption, so a statement's chunk names the report, the
+unit and the period. Nested tables and the tables of other parsers still take the converter's
+path with the first row promoted to header. Measured on the NVDA earnings release: 52 chunks
+became 32, none of them a table without its header.
+
**Origin and anchor.** `chunks.origin` (stated, ocr, transcribed, described), `origin_model`, and an
`anchor` whose shape is checked per origin (page and box; start, end and speakers; page and image or
part); the packer never mixes origins in one chunk; a described block is its own chunk with the
@@ -85,8 +101,8 @@ indexed and never extracted (`sources.config.extract`) [0036 d7].
anchor in the interface, opening the page, image or recording [0040].
- Feishu and Confluence connectors with a block-tree renderer and sampled versions [0013].
- An external parser behind the block model for PDF layout; evidence naming a table cell or an image
- region; real Markdown tables from the DOCX, PDF, spreadsheet and CSV parsers; header rows promoted
- by content [0039].
+ region; real Markdown tables from the DOCX, PDF, spreadsheet and CSV parsers, rendered the way
+ HTML tables now are [0039].
- Backup and restore (#712), a 100k-document benchmark (#713) [0016 E].
## Open questions