Mova · papers by stage
tree → frames
Reading the tree into frames — who did what to whom — with speech acts, hedges and coreference. The 20 works of Mova's own list that this stage uses; measured marks a work whose effect Mova measured in its own experiments (8 here). Each title opens our card: the summary, how Mova uses the work, its references and who cites it. What the stage does is on Mova's pipeline page; every stage is on the main list.
- Resolving Pronoun ReferencesmeasuredJerry R. Hobbs · 1978
What Mova took:
coref/src/sieve.rs— the candidate order for pronoun resolution follows the spirit of Hobbs: breadth-first, left-to-right tree traversal ordered by depth of the head and then position; first the mention's own clause, then the… - A Model-Theoretic Coreference Scoring SchememeasuredMarc Vilain et al. · 1995
What Mova took:
coref/src/score.rs: the MUC metric is implemented in Rust as one of the coreference scores, alongside B³, CEAF-e/CEAF-m, and CoNLL F1 (the average of MUC, B³ and CEAF-e). - Algorithms for Scoring Coreference ChainsmeasuredAmit Bagga, Breck Baldwin · 1998
What Mova took:
coref/src/score.rs— the coreference scorer implements B³ alongside MUC, CEAF-e and CEAF-m; the CoNLL F1 reported for Mova's coreference resolver is the mean of MUC, B³ and CEAF-e. - Hedging in Scientific Research ArticlesmeasuredKen Hyland · 1998
What Mova took:
prag/data/cues.tsv— a small hand-made list of pragmatic cue words and phrases; the hedge entries (maybe, perhaps, probably, possibly, apparently, I think, I suppose, I guess…) are compiled with this book and Lakoff 1973 as… - On Coreference Resolution Performance MetricsmeasuredXiaoqiang Luo · 2005
What Mova took:
coref/src/score.rs— Mova's own Rust scorer implements CEAF-e (φ4, entity-based) and CEAF-m (φ3, mention-based), with the optimal entity alignment found by Kuhn-Munkres, computed separately per connected component of the… - A Multi-Pass Sieve for Coreference ResolutionmeasuredKarthik Raghunathan et al. · 2010
What Mova took:
coref/src/sieve.rs— the overall architecture: an ordered list of sieves from most to least precise (speaker, exact, relaxed, appos, pred, relpron, acronym, reflexive, strict head-match variants, proper_head, pronoun), each… - Deterministic Coreference Resolution Based on Entity-Centric, Precision-Ranked RulesmeasuredHeeyoung Lee et al. · 2013
What Mova took:
coref/src/sieve.rs— overall architecture: an ordered list of sieves from most to least precise… - Scoring Coreference Partitions of Predicted Mentions: A Reference ImplementationmeasuredSameer Pradhan et al. · 2014
What Mova took:
coref/src/score.rs— the MUC, B-cubed, CEAF-e and CEAF-m formulas for predicted mentions follow this paper: no mention manipulation, unmatched mentions just do not contribute; CoNLL F1 is the mean of MUC, B-cubed and CEAF-e… - Speech Acts: An Essay in the Philosophy of LanguageJohn R. Searle · 1969
What Mova took:
prag/data/cues.tsv— the cue list's markers of illocutionary force (thanks, apologies, greetings, promises, warnings, suggestions, offers, yes/no) are compiled with this book (and Searle 1975 on indirect speech acts) as the… - Hedges: A Study in Meaning Criteria and the Logic of Fuzzy ConceptsGeorge Lakoff · 1973
What Mova took:
prag/data/cues.tsv— the hedge entries in Mova's small hand-made cue list (sort of, kind of, maybe, I think and similar) take this paper, together with Hyland 1998, as their conceptual basis; the list itself is Mova's… - Politeness: Some Universals in Language UsagePenelope Brown, Stephen C. Levinson · 1987
What Mova took:
prag/data/cues.tsv— a small hand-written cue list for the pragmatics crate; politeness markers (e.g. "please", apology and thanks cues) are grounded in this book, alongside Lakoff 1973 and Hyland 1998 for hedges and Searle for… - An Argument for Basic EmotionsPaul Ekman · 1992
What Mova took:
global/seeds/shortcuts/feelings.md: the seed that maps event types to feelings… - An Algorithm for Pronominal Anaphora ResolutionShalom Lappin, Herbert J. Leass · 1994
What Mova took: the development notes of
coref("What next") — the salience weights of this paper… - CogNIAC: A Discourse Processing EngineBreck Baldwin · 1995
What Mova took:
coref/src/sieve.rs(find_quote_authors) — determining who speaks in a quotation: in a sentence containing a quote, the author is the subject of a speech verb outside the quotation marks; speech verbs come from a closed… - CogNIAC: high precision coreference with limited knowledge and linguistic resourcesBreck Baldwin · 1997
What Mova took: the development notes of
coref, sievespeaker: the rules for first- and second-person pronouns within one voice… - Simple Coreference Resolution with Rich Syntactic and Semantic FeaturesAria Haghighi, Dan Klein · 2009
What Mova took: the development notes of
corefandcoref/src/sieve.rs: Mova's coreference module is a deterministic multi-pass sieve (architecture after Raghunathan et al. 2010 and Lee et al. 2013). From this paper it takes the principle… - Udapi: Universal API for Universal DependenciesMartin Popel, Zdeněk Žabokrtský, Martin Vojtek · 2017
What Mova took:
coref/src/score.rsandcoref/src/corefud.rs— the head of a mention span is computed as in Udapi'scorefud.MoveHeadblock: the word whose own head lies outside the span; among several candidates, the highest in the tree… - CorefUD 1.0: Coreference Meets Universal DependenciesAnna Nedoluzhko et al. · 2022
What Mova took:
coref—coref resolvewrites its output in the CorefUD format:Entity=annotations in MISC with a# global.Entity = eid-etype-head-otherheader, whereheadis the position of the head inside the span andotherrecords… - Findings of the Fourth Shared Task on Multilingual Coreference Resolution: Can LLMs Dethrone Traditional Approaches?Michal Novák et al. · 2025
What Mova took:
coref/src/score.rs— mention matching by head, as in the shared task's primary metric: gold and system mentions match when they share a head; several mentions with the same head are disambiguated by span… - Findings of the Fifth Shared Task on Multilingual Coreference Resolution: Expanding Datasets for Long-Range EntitiesMichal Novák et al. · 2026
What Mova took:
coref/src/score.rs— together with the 2025 edition, the source for head-based mention matching as the primary scoring mode (CRAC 2023–2026 convention).