---
title: "Integrating Predictions from Neural-Network — Roesiger 2018"
description: "Resolving coreference and bridging anaphora often depends on lexical-semantic knowledge, such as knowing that a room has windows or that malaria is a…"
authors: ["Ina Roesiger", "Maximilian Köper", "Kim Anh Nguyen", "Sabine Schulte im Walde"]
year: 2018
venue: "Proceedings of the First Workshop on Computational Models of Reference, Anaphora and Coreference"
source: "https://aclanthology.org/W18-0705/"
doi: "10.18653/v1/w18-0705"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/papers/roesiger-2018-integrating-predictions-neural-network-relation/"
---
[Elamur Research](https://md.research.elamur.ai/) · [this page in HTML](https://research.elamur.ai/free/papers/roesiger-2018-integrating-predictions-neural-network-relation/)

[home](https://md.research.elamur.ai/) / [papers](https://md.research.elamur.ai/free/papers/) / [2018](https://md.research.elamur.ai/free/papers/2018/page/2/) / Integrating Predictions from Neural-Network Relation Classifiers into Coreference and Bridging Resolution

# Integrating Predictions from Neural-Network Relation Classifiers into Coreference and Bridging Resolution

**Authors:** [Ina Roesiger](https://md.research.elamur.ai/free/authors/ina-roesiger/), Maximilian Köper, Kim Anh Nguyen, [Sabine Schulte im Walde](https://md.research.elamur.ai/free/authors/sabine-schulte-im-walde/) · **Year:** 2018 · **Venue:** Proceedings of the First Workshop on Computational Models of Reference, Anaphora and Coreference · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/W18-0705/) · [PDF](https://aclanthology.org/W18-0705.pdf) · [DOI 10.18653/v1/w18-0705](https://doi.org/10.18653/v1/w18-0705) · [OpenAlex](https://openalex.org/W2806291270)

**Citations:** 8 citations (openalex-snapshot · 2026-10-14) · [5](#cited-by) in our library, of the 140673 papers whose bibliographies we read

## Summary
Resolving coreference and bridging anaphora often depends on lexical-semantic knowledge, such as knowing that a room has windows or that malaria is a disease. Such knowledge usually comes from hand-built resources like WordNet. The authors instead test whether relation likelihoods predicted by a neural classifier can serve this purpose. They train feed-forward networks on noun-pair relation benchmarks and score candidate anaphor–antecedent head pairs drawn from OntoNotes and ISNotes.

The predictions are used in two ways:
- as features for the IMS HotCoref coreference system
- as a new rule in a re-implementation of the rule-based bridging resolver by Hou et al. (2014)

Two setups are compared. The first uses BLESS with ConceptNet vectors. The second uses WordNet-derived relations with fastText embeddings, which is more balanced and covers more vocabulary.

Bridging resolution gains modestly, with meronymy combined with a cosine threshold working best. Coreference does not improve, since its baseline is already strong. The study suggests that automatically induced relations are useful mainly for the low-performing bridging task. It also notes that real corpus pairs show messier relation patterns than the authors expected.

## Key points
- **Data:** about 65,113 unique coreference pairs (OntoNotes) and 633 bridging pairs (ISNotes) were scored.
- **Experiment 1 classifier:** BLESS with 26,546 pairs and 300-dimensional ConceptNet vectors, fed to a network with two hidden layers (200 and 150 units).
  - Accuracy was 87.8% on the standard split.
  - Accuracy dropped to 58.6% once lexical overlap between train and test was removed.
- **Bridging result:** a new rule for unmodified "the N" anaphors that requires meronymy and cosine > 0.2 raised F1 from 18.0 to 20.38, a significant gain.
- **Experiment 2:** WordNet relations (2,010 sampled pairs per relation) and 100-dimensional fastText vectors trained on DECOW14.
  - Accuracy was 55.8%.
  - Bridging improved only with meronymy at cosine > 0.3, reaching F1 18.92.
- **Coreference:** adding the relation features to IMS HotCoref lowered the CoNLL score in every configuration.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

Fields: [Linguistics & languages](https://md.research.elamur.ai/free/shelves/linguistics/)

Topics: [Coreference resolution](https://md.research.elamur.ai/free/topics/coreference-resolution/)

## Cited by

Cited by

5 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 3 not yet.

- [Bridging Resolution: A Survey of the State of the Art](https://md.research.elamur.ai/free/papers/kobayashi-2020-bridging-resolution-survey-state-art/)
- [Bridging resolution: Task definition, corpus resources and rule-based experiments](https://md.research.elamur.ai/free/papers/roesiger-2018-bridging-resolution-task-definition-corpus/)

## Cite

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{roesiger2018integrating,
  title = {{Integrating Predictions from Neural-Network Relation Classifiers into Coreference and Bridging Resolution}},
  author = {Roesiger, Ina and Köper, Maximilian and Nguyen, Kim Anh and Walde, Sabine Schulte im},
  booktitle = {Proceedings of the First Workshop on Computational Models of Reference, Anaphora and Coreference},
  year = {2018},
  doi = {10.18653/v1/w18-0705},
  url = {https://doi.org/10.18653/v1/w18-0705}
}
```

- [← previous Improving the Transformer Translation Model with Document-Level Context](https://md.research.elamur.ai/free/papers/zhang-2018-improving-transformer-translation-model-document/)
- [next → Integrating Stance Detection and Fact Checking in a Unified Corpus](https://md.research.elamur.ai/free/papers/baly-2018-integrating-stance-detection-fact-checking/)
- [↑ back to papers of 2018](https://md.research.elamur.ai/free/papers/2018/page/2/)

## References

23 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Vincent Ng. 2010. Supervised noun phrase coreference research: The first fifteen years. In Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics, pages 1396–1411, Uppsala, Sweden. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Vincent+Ng%2E+2010%2E+Supervised+noun+phrase+coreference+research%3A+The+first+fifteen+years%2E+In+Proceedings+of+the+48th+Annual+Meeting+of+the+Association+for+Computational+Linguistics%2C+pages+1396%E2%80%931411%2C+Upp) [our card](https://md.research.elamur.ai/free/papers/ng-2010-supervised-noun-phrase-coreference-research/)
2. Marco Baroni and Alessandro Lenci. 2011. How we blessed distributional semantic evaluation. In Proceedings of the GEMS 2011 Workshop on GEometrical Models of Natural Language Semantics, GEMS ’11, pages 1–10, Stroudsburg, PA, USA. [find](https://scholar.google.com/scholar?q=Marco+Baroni+and+Alessandro+Lenci%2E+2011%2E+How+we+blessed+distributional+semantic+evaluation%2E+In+Proceedings+of+the+GEMS+2011+Workshop+on+GEometrical+Models+of+Natural+Language+Semantics%2C+GEMS+%E2%80%9911%2C+page)
3. Massimo Poesio and Ron Artstein. 2005. The reliability of anaphoric annotation, reconsidered: Taking ambiguity into account. In Proceedings of the workshop on frontiers in corpus annotations ii: Pie in the sky, pages 76–83. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Massimo+Poesio+and+Ron+Artstein%2E+2005%2E+The+reliability+of+anaphoric+annotation%2C+reconsidered%3A+Taking+ambiguity+into+account%2E+In+Proceedings+of+the+workshop+on+frontiers+in+corpus+annotations+ii%3A+Pie+i) [our card](https://md.research.elamur.ai/free/papers/poesio-2005-reliability-anaphoric-annotation-reconsidered-taking/)
4. Anders Björkelund and Jonas Kuhn. 2014. Learning structured perceptrons for coreference resolution with latent antecedents and non-local features. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 47–57, Baltimore, Maryland. As… [find](https://scholar.google.com/scholar?q=Anders+Bjo%CC%88rkelund+and+Jonas+Kuhn%2E+2014%2E+Learning+structured+perceptrons+for+coreference+resolution+with+latent+antecedents+and+non%2Dlocal+features%2E+In+Proceedings+of+the+52nd+Annual+Meeting+of+the+Ass) [our card](https://md.research.elamur.ai/free/papers/bjorkelund-2014-learning-structured-perceptrons-coreference-resolution/)
5. Massimo Poesio, Tomonori Ishikawa, Sabine Schulte Im Walde, and Renata Vieira. 2002. Acquiring lexical knowledge for anaphora resolution. In Proceedings of the third international conference on language resources and evaluation (LREC), Las Palmas, Spain. [find](https://scholar.google.com/scholar?q=Massimo+Poesio%2C+Tomonori+Ishikawa%2C+Sabine+Schulte+Im+Walde%2C+and+Renata+Vieira%2E+2002%2E+Acquiring+lexical+knowledge+for+anaphora+resolution%2E+In+Proceedings+of+the+third+international+conference+on+langua) [our card](https://md.research.elamur.ai/free/papers/poesio-2002-acquiring-lexical-knowledge-anaphora-resolution/)
6. Piotr Bojanowski, Edouard Grave, Armand Joulin, and Tomas Mikolov. 2016. Enriching word vectors with subword information. arXiv preprint arXiv:1607.04606. [find](https://scholar.google.com/scholar?q=Piotr+Bojanowski%2C+Edouard+Grave%2C+Armand+Joulin%2C+and+Tomas+Mikolov%2E+2016%2E+Enriching+word+vectors+with+subword+information%2E+arXiv+preprint+arXiv%3A1607%2E04606%2E)
7. Massimo Poesio and Renata Vieira. 1998. A corpus-based investigation of definite description use. Computational Linguistics, 24(2):183–216. [find](https://scholar.google.com/scholar?q=Massimo+Poesio+and+Renata+Vieira%2E+1998%2E+A+corpus%2Dbased+investigation+of+definite+description+use%2E+Computational+Linguistics%2C+24%282%29%3A183%E2%80%93216%2E)
8. Kevin Clark and Christopher D. Manning. 2016. Deep reinforcement learning for mention-ranking coreference models. In Proceedings of the 2016 Conference on Empirical Methods on Natural Language Processing, Austin, USA. [find](https://scholar.google.com/scholar?q=Kevin+Clark+and+Christopher+D%2E+Manning%2E+2016%2E+Deep+reinforcement+learning+for+mention%2Dranking+coreference+models%2E+In+Proceedings+of+the+2016+Conference+on+Empirical+Methods+on+Natural+Language+Process) [our card](https://md.research.elamur.ai/free/papers/clark-2016-deep-reinforcement-learning-mention-ranking/)
9. Simone Paolo Ponzetto and Michael Strube. 2007. Knowledge derived from wikipedia for computing semantic relatedness. J. Artif. Intell. Res.(JAIR), 30:181–212. [find](https://scholar.google.com/scholar?q=Simone+Paolo+Ponzetto+and+Michael+Strube%2E+2007%2E+Knowledge+derived+from+wikipedia+for+computing+semantic+relatedness%2E+J%2E+Artif%2E+Intell%2E+Res%2E%28JAIR%29%2C+30%3A181%E2%80%93212%2E)
10. Greg Durrett and Dan Klein. 2013. Easy victories and uphill battles in coreference resolution. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1971–1982, Seattle, USA. [find](https://scholar.google.com/scholar?q=Greg+Durrett+and+Dan+Klein%2E+2013%2E+Easy+victories+and+uphill+battles+in+coreference+resolution%2E+In+Proceedings+of+the+2013+Conference+on+Empirical+Methods+in+Natural+Language+Processing%2C+pages+1971%E2%80%93198) [our card](https://md.research.elamur.ai/free/papers/durrett-2013-easy-victories-uphill-battles-coreference/)
11. Sameer Pradhan, Xiaoqiang Luo, Marta Recasens, Eduard Hovy, Vincent Ng, and Michael Strube. 2014. Scoring coreference partitions of predicted mentions: A reference implementation. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers), pag… [find](https://scholar.google.com/scholar?q=Sameer+Pradhan%2C+Xiaoqiang+Luo%2C+Marta+Recasens%2C+Eduard+Hovy%2C+Vincent+Ng%2C+and+Michael+Strube%2E+2014%2E+Scoring+coreference+partitions+of+predicted+mentions%3A+A+reference+implementation%2E+In+Proceedings+of+th) [our card](https://md.research.elamur.ai/free/papers/pradhan-2014-scoring-coreference-partitions-predicted-mentions/)
12. Christiane Fellbaum. 1998. WordNet: An Electronic Lexical Database. Bradford Books. [find](https://scholar.google.com/scholar?q=Christiane+Fellbaum%2E+1998%2E+WordNet%3A+An+Electronic+Lexical+Database%2E+Bradford+Books%2E)
13. Tim Feuerbach, Martin Riedl, and Chris Biemann. 2015. Distributional semantics for resolving bridging mentions. In Proceedings of the International Conference Recent Advances in Natural Language Processing, pages 192–199, Hissar, Bulgaria. [find](https://scholar.google.com/scholar?q=Tim+Feuerbach%2C+Martin+Riedl%2C+and+Chris+Biemann%2E+2015%2E+Distributional+semantics+for+resolving+bridging+mentions%2E+In+Proceedings+of+the+International+Conference+Recent+Advances+in+Natural+Language+Proce)
14. Roland Schäfer. 2015. Processing and Querying Large Web Corpora with the COW14 Architecture. In Proceedings of the 3rd Workshop on Challenges in the Management of Large Corpora, pages 28 – 34. [find](https://scholar.google.com/scholar?q=Roland+Scha%CC%88fer%2E+2015%2E+Processing+and+Querying+Large+Web+Corpora+with+the+COW14+Architecture%2E+In+Proceedings+of+the+3rd+Workshop+on+Challenges+in+the+Management+of+Large+Corpora%2C+pages+28+%E2%80%93+34%2E)
15. Yufang Hou. 2016. Unrestricted Bridging Resolution. Ph.D. thesis. [find](https://scholar.google.com/scholar?q=Yufang+Hou%2E+2016%2E+Unrestricted+Bridging+Resolution%2E+Ph%2ED%2E+thesis%2E) [our card](https://md.research.elamur.ai/free/papers/hou-2018-unrestricted-bridging-resolution/)
16. Vered Shwartz and Ido Dagan. 2016. Path-based vs. distributional information in recognizing lexical semantic relations. In Proceedings of the 5th Workshop on Cognitive Aspects of the Lexicon (CogALex-V), in COLING, Osaka, Japan. [find](https://scholar.google.com/scholar?q=Vered+Shwartz+and+Ido+Dagan%2E+2016%2E+Path%2Dbased+vs%2E+distributional+information+in+recognizing+lexical+semantic+relations%2E+In+Proceedings+of+the+5th+Workshop+on+Cognitive+Aspects+of+the+Lexicon+%28CogALex%2D)
17. Yufang Hou, Katja Markert, and Michael Strube. 2014. A rule-based system for unrestricted bridging resolution: Recognizing bridging anaphora and finding links to antecedents. In Proceedings of the Conference on Empirical Methods in Natural Language Processing, pages 2082–2093, Seattle, USA. [find](https://scholar.google.com/scholar?q=Yufang+Hou%2C+Katja+Markert%2C+and+Michael+Strube%2E+2014%2E+A+rule%2Dbased+system+for+unrestricted+bridging+resolution%3A+Recognizing+bridging+anaphora+and+finding+links+to+antecedents%2E+In+Proceedings+of+the+Con) [our card](https://md.research.elamur.ai/free/papers/hou-2014-rule-based-system-unrestricted-bridging/)
18. Sidney Siegel and N. John Jr. Castellan. 1988. Non-parametric Statistics for the Behavioral Sciences, 2nd edition. McGraw-Hill, Berkeley, CA. [find](https://scholar.google.com/scholar?q=Sidney+Siegel+and+N%2E+John+Jr%2E+Castellan%2E+1988%2E+Non%2Dparametric+Statistics+for+the+Behavioral+Sciences%2C+2nd+edition%2E+McGraw%2DHill%2C+Berkeley%2C+CA%2E)
19. Robert Speer, Joshua Chin, and Catherine Havasi. 2017. Conceptnet 5.5: An open multilingual graph of general knowledge. In AAAI. [find](https://scholar.google.com/scholar?q=Robert+Speer%2C+Joshua+Chin%2C+and+Catherine+Havasi%2E+2017%2E+Conceptnet+5%2E5%3A+An+open+multilingual+graph+of+general+knowledge%2E+In+AAAI%2E)
20. Omer Levy, Steffen Remus, Chris Biemann, and Ido Dagan. 2015. Do supervised distributional methods really learn lexical inference relations? In Proceedings of the 2015 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 970–97… [find](https://scholar.google.com/scholar?q=Omer+Levy%2C+Steffen+Remus%2C+Chris+Biemann%2C+and+Ido+Dagan%2E+2015%2E+Do+supervised+distributional+methods+really+learn+lexical+inference+relations%3F+In+Proceedings+of+the+2015+Conference+of+the+North+American) [our card](https://md.research.elamur.ai/free/papers/levy-2015-do-supervised-distributional-methods-really/)
21. Renata Vieira and Massimo Poesio. 2000. An empirically based system for processing definite descriptions. Computational Linguistics, 26(4):539–593. [find](https://scholar.google.com/scholar?q=Renata+Vieira+and+Massimo+Poesio%2E+2000%2E+An+empirically+based+system+for+processing+definite+descriptions%2E+Computational+Linguistics%2C+26%284%29%3A539%E2%80%93593%2E)
22. Ralph Weischedel, Sameer Pradhan, Lance Ramshaw, Martha Palmer, Nianwen Xue, Mitchell Marcus, Ann Taylor, Craig Greenberg, Eduard Hovy, Robert Belvin, et al. 2011. Ontonotes release 4.0. LDC2011T03, Philadelphia, Penn.: Linguistic Data Consortium. [find](https://scholar.google.com/scholar?q=Ralph+Weischedel%2C+Sameer+Pradhan%2C+Lance+Ramshaw%2C+Martha+Palmer%2C+Nianwen+Xue%2C+Mitchell+Marcus%2C+Ann+Taylor%2C+Craig+Greenberg%2C+Eduard+Hovy%2C+Robert+Belvin%2C+et+al%2E+2011%2E+Ontonotes+release+4%2E0%2E+LDC2011T03%2C+P)
23. Katja Markert, Yufang Hou, and Michael Strube. 2012. Collective classification for fine-grained information status. In Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics: Long Papers-Volume 1, pages 795–804, Jeju Island, Korea. Association for Computational Lingu… [find](https://scholar.google.com/scholar?q=Katja+Markert%2C+Yufang+Hou%2C+and+Michael+Strube%2E+2012%2E+Collective+classification+for+fine%2Dgrained+information+status%2E+In+Proceedings+of+the+50th+Annual+Meeting+of+the+Association+for+Computational+Lingu) [our card](https://md.research.elamur.ai/free/papers/markert-2012-collective-classification-fine-grained-information/)