---
title: "Exploiting Argument Information to Improve Event — Liu 2017"
description: "The paper addresses event detection (ED) on ACE 2005: finding trigger words and assigning them one of 33 event types."
authors: ["Shulin Liu", "Yubo Chen", "Kang Liu", "Jun Zhao"]
year: 2017
venue: "Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/P17-1164/"
doi: "10.18653/v1/p17-1164"
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/en/papers/liu-2017-exploiting-argument-information-improve-event/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/liu-2017-exploiting-argument-information-improve-event/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2017](https://md.research.elamur.ai/free/en/papers/2017/page/2/) / Exploiting Argument Information to Improve Event Detection via Supervised Attention Mechanisms

# Exploiting Argument Information to Improve Event Detection via Supervised Attention Mechanisms

**Authors:** [Shulin Liu](https://md.research.elamur.ai/free/en/authors/shulin-liu/), [Yubo Chen](https://md.research.elamur.ai/free/en/authors/yubo-chen/), [Kang Liu](https://md.research.elamur.ai/free/en/authors/kang-liu/), [Jun Zhao](https://md.research.elamur.ai/free/en/authors/jun-zhao/) · **Year:** 2017 · **Venue:** Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/P17-1164/) · [PDF](https://aclanthology.org/P17-1164.pdf) · [DOI 10.18653/v1/p17-1164](https://doi.org/10.18653/v1/p17-1164) · [OpenAlex](https://openalex.org/W2739918945)

**Citations:** 237 citations (openalex-snapshot · 2026-10-14) · [72](#cited-by) in our library, of the 140884 papers whose bibliographies we read

## Summary
The paper addresses event detection (ED) on ACE 2005: finding trigger words and assigning them one of 33 event types. The authors observe that existing systems either ignore annotated event arguments or use them only indirectly. Joint ED+AE models mostly improve argument extraction and barely help ED. Their explanations are that arguments outnumber triggers in the training data, and that trigger pre-prediction in joint models uses no argument information. They propose a neural classifier that labels each token, using attention over context words and their entity types. The attention is trained with direct supervision. Gold attention distributions are built from the annotated arguments, so the model learns to look at argument words when judging a candidate trigger. At test time no argument annotation is needed. Two strategies for building gold attention are compared. Optionally, FrameNet-derived events are added as extra training data. The best variant reaches 71.7 F1 on ACE 2005, or 71.9 with FrameNet data. This beats prior state-of-the-art results with notably higher precision, showing that arguments can help ED when used explicitly.

## Key points
- Table 1 compares joint and pipeline systems. Joint systems gain over 1.9 points on AE but under 0.2 on ED. Li's joint symbolic model is even worse on ED than Hong's pipeline (67.5 vs. 68.3).
- ACE 2005 has roughly 9800 annotated arguments versus 5300 triggers. The authors suggest this imbalance pushes joint models to favour AE.
- ED is framed as 34-way token classification (33 types plus NA). Inputs are Skip-gram embeddings trained on the NYT corpus and randomly initialised entity-type embeddings.
- Results on ACE 2005:
  - Plain unsupervised attention (ANN) reaches 65.0 F1.
  - ANN-S1 reaches 70.8 F1.
  - ANN-S2 reaches 71.7 F1 (P 78.0, R 66.3). S2 also attends to words surrounding the arguments, which trades a little precision for recall.
  - A t-test (p ≤ 0.05) shows ANN-S2 significantly outperforms the compared methods.
- FrameNet augmentation reuses events released by Liu et al. (2016a):
  - Core frame elements are treated as arguments.
  - ACE entities are labelled with the RPI Joint Information Extraction System.
  - ANN-S2 rises to 71.9 F1, and plain ANN to 66.7.

## Interesting ideas
- **Supervising attention with train-only annotations.** Turn an annotation that exists only in training data (here, argument spans) into a target distribution for attention. This injects structural knowledge without needing that annotation at inference, and could transfer to tasks such as relation or sentiment classification.
- **Diagnosing why joint learning fails to help a subtask.** The authors check label-count imbalance between subtasks and which components actually consume the other subtask's signal. This is a useful audit before assuming multi-task training helps every task.
- **Separate attention over words and entity types.** Each attention is computed against the candidate word. The candidate's own entity type is left out because few true triggers are entities. This reflects a thoughtful choice about which features should drive attention.
- **Tuning the gold-attention target as a precision/recall knob.** A tight target on argument words alone versus a looser one that includes neighbouring words lets the designer shift between precision and recall.
- **Repurposing a lexical-semantic resource as silver training data.** Frame elements in FrameNet are mapped to event arguments, and missing entity layers are filled with an off-the-shelf IE tool. This could inspire similar augmentation for low-resource languages, for example where frame-like resources exist for Ukrainian.

*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: [Natural language processing](https://md.research.elamur.ai/free/en/shelves/nlp/)

Topics: [Semantic role labeling](https://md.research.elamur.ai/free/en/topics/semantic-role-labeling/)

## Cited by

Cited by

72 papers of our library cite it (of the 140884 whose bibliographies we read); 5 of them have a page here, 67 not yet.

- [Cross-Document Event Coreference Resolution on Discourse Structure](https://md.research.elamur.ai/free/en/papers/chen-2023-cross-document-event-coreference-resolution/)
- [Jointly Multiple Events Extraction via Attention-based Graph Information Aggregation](https://md.research.elamur.ai/free/en/papers/liu-2018-jointly-multiple-events-extraction-attention/)
- [Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing](https://md.research.elamur.ai/free/en/papers/liu-2021-pre-train-prompt-predict-systematic/)
- [MAVEN: A Massive General Domain Event Detection Dataset](https://md.research.elamur.ai/free/en/papers/wang-2020-maven-massive-general-domain-event/)
- [Exploring Pre-trained Language Models for Event Extraction and Generation](https://md.research.elamur.ai/free/en/papers/yang-2019-exploring-pre-trained-language-models/)

## Cite

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

```bibtex
@inproceedings{liu2017exploiting,
  title = {{Exploiting Argument Information to Improve Event Detection via Supervised Attention Mechanisms}},
  author = {Liu, Shulin and Chen, Yubo and Liu, Kang and Zhao, Jun},
  booktitle = {Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2017},
  doi = {10.18653/v1/p17-1164},
  url = {https://doi.org/10.18653/v1/p17-1164}
}
```

- [← previous Explanation in Artificial Intelligence: Insights from the Social Sciences](https://md.research.elamur.ai/free/en/papers/miller-2017-explanation-artificial-intelligence-insights-social/)
- [next → Exploiting Cross-Sentence Context for Neural Machine Translation](https://md.research.elamur.ai/free/en/papers/wang-2017-exploiting-cross-sentence-context-neural/)
- [↑ back to papers of 2017](https://md.research.elamur.ai/free/en/papers/2017/page/2/)

## References

28 references, 12 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. David Ahn. 2006. Proceedings of the workshop on annotating and reasoning about time and events. Association for Computational Linguistics, pages 1–8. [http://aclweb.org/anthology/W06-0901](http://aclweb.org/anthology/W06-0901). [↗](http://aclweb.org/anthology/W06-0901) [our card](https://md.research.elamur.ai/free/en/papers/ahn-2006-stages-event-extraction/)
2. Yoshua Bengio, Réjean Ducharme, Pascal Vincent, and Christian Janvin. 2003. A neural probabilistic language model. The Journal of Machine Learning Research 3:1137–1155. [find](https://scholar.google.com/scholar?q=Yoshua+Bengio%2C+Re%CC%81jean+Ducharme%2C+Pascal+Vincent%2C+and+Christian+Janvin%2E+2003%2E+A+neural+probabilistic+language+model%2E+The+Journal+of+Machine+Learning+Research+3%3A1137%E2%80%931155%2E)
3. Yubo Chen, Liheng Xu, Kang Liu, Daojian Zeng, and Jun Zhao. 2015. Event extraction via dynamic multi-pooling convolutional neural networks. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Proces… [find](https://scholar.google.com/scholar?q=Yubo+Chen%2C+Liheng+Xu%2C+Kang+Liu%2C+Daojian+Zeng%2C+and+Jun+Zhao%2E+2015%2E+Event+extraction+via+dynamic+multi%2Dpooling+convolutional+neural+networks%2E+In+Proceedings+of+the+53rd+Annual+Meeting+of+the+Association) [our card](https://md.research.elamur.ai/free/en/papers/chen-2015-event-extraction-dynamic-multi-pooling/)
4. Dumitru Erhan, Yoshua Bengio, Aaron Courville, Pierre-Antoine Manzagol, Pascal Vincent, and Samy Bengio. 2010. Why does unsupervised pre-training help deep learning? The Journal of Machine Learning Research 11:625–660. [find](https://scholar.google.com/scholar?q=Dumitru+Erhan%2C+Yoshua+Bengio%2C+Aaron+Courville%2C+Pierre%2DAntoine+Manzagol%2C+Pascal+Vincent%2C+and+Samy+Bengio%2E+2010%2E+Why+does+unsupervised+pre%2Dtraining+help+deep+learning%3F+The+Journal+of+Machine+Learning+Re)
5. Collin F. Baker, Charles J. Fillmore, and John B. Lowe. 1998. The berkeley framenet project. In COLING 1998 Volume 1: The 17th International Conference on Computational Linguistics. [http://aclweb.org/anthology/C98-1013](http://aclweb.org/anthology/C98-1013). [↗](http://aclweb.org/anthology/C98-1013)
6. Prashant Gupta and Heng Ji. 2009. Predicting unknown time arguments based on cross-event propagation. In Proceedings of the ACL-IJCNLP 2009 Conference Short Papers. Association for Computational Linguistics, pages 369– 372. [http://aclweb.org/anthology/P09-2093](http://aclweb.org/anthology/P09-2093). [↗](http://aclweb.org/anthology/P09-2093)
7. Martin T Hagan, Howard B Demuth, Mark H Beale, et al. 1996. Neural network design. Pws Pub. Boston. [find](https://scholar.google.com/scholar?q=Martin+T+Hagan%2C+Howard+B+Demuth%2C+Mark+H+Beale%2C+et+al%2E+1996%2E+Neural+network+design%2E+Pws+Pub%2E+Boston%2E)
8. Geoffrey E Hinton, Nitish Srivastava, Alex Krizhevsky, Ilya Sutskever, and Ruslan R Salakhutdinov. 2012. Improving neural networks by preventing co-adaptation of feature detectors. arXiv preprint arXiv:1207.0580 https://arxiv.org/abs/1207.0580. [↗](https://arxiv.org/abs/1207.0580)
9. Yu Hong, Jianfeng Zhang, Bin Ma, Jianmin Yao, Guodong Zhou, and Qiaoming Zhu. 2011. Using cross-entity inference to improve event extraction. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies. Association for Computational Linguis… [find](https://scholar.google.com/scholar?q=Yu+Hong%2C+Jianfeng+Zhang%2C+Bin+Ma%2C+Jianmin+Yao%2C+Guodong+Zhou%2C+and+Qiaoming+Zhu%2E+2011%2E+Using+cross%2Dentity+inference+to+improve+event+extraction%2E+In+Proceedings+of+the+49th+Annual+Meeting+of+the+Associati) [our card](https://md.research.elamur.ai/free/en/papers/hong-2011-cross-entity-inference-improve-event/)
10. Heng Ji and Ralph Grishman. 2008. Refining event extraction through cross-document inference. In Proceedings of ACL-08: HLT . Association for Computational Linguistics, pages 254–262. [http://aclweb.org/anthology/P08-1030](http://aclweb.org/anthology/P08-1030). [↗](http://aclweb.org/anthology/P08-1030) [our card](https://md.research.elamur.ai/free/en/papers/ji-2008-refining-event-extraction-through-cross/)
11. Yoon Kim. 2014. Convolutional neural networks for sentence classification. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing. pages 1746–1751. [http://www.anthology.aclweb.org/D14-1181](http://www.anthology.aclweb.org/D14-1181). [↗](http://www.anthology.aclweb.org/D14-1181) [our card](https://md.research.elamur.ai/free/en/papers/kim-2014-convolutional-neural-networks-sentence-classification/)
12. Qi Li and Heng Ji. 2014. Incremental joint extraction of entity mentions and relations. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, pages 402–412. https://doi.org/10.3115/v1/P14-1038. [↗](https://doi.org/10.3115/v1/P14-1038) [our card](https://md.research.elamur.ai/free/en/papers/li-2014-incremental-joint-extraction-entity-mentions/)
13. Qi Li, Heng Ji, Yu HONG, and Sujian Li. 2014. Constructing information networks using one single model. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computational Linguistics, pages 1846–1851. https://doi.org/10.3115/v1/D14-1198. [↗](https://doi.org/10.3115/v1/D14-1198)
14. Qi Li, Heng Ji, and Liang Huang. 2013. Joint event extraction via structured prediction with global features. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, pages 73–82. [http://aclweb.org/…](http://aclweb.org/…) [↗](http://aclweb.org/…) [our card](https://md.research.elamur.ai/free/en/papers/li-2013-joint-event-extraction-structured-prediction/)
15. Shasha Liao and Ralph Grishman. 2010. Using document level cross-event inference to In Proceedings improve event extraction. of the 48th Annual Meeting of the Association for Computational Linguistics. Association for Computational Linguistics, pages 789–797. [http://aclweb.org/anthology/P10-1081](http://aclweb.org/anthology/P10-1081). [↗](http://aclweb.org/anthology/P10-1081) [our card](https://md.research.elamur.ai/free/en/papers/liao-2010-document-level-cross-event-inference/)
16. Shulin Liu, Yubo Chen, Shizhu He, Kang Liu, and Jun Zhao. 2016a. Leveraging framenet to improve automatic event detection. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistics, pages 2134–2143. h… [find](https://scholar.google.com/scholar?q=Shulin+Liu%2C+Yubo+Chen%2C+Shizhu+He%2C+Kang+Liu%2C+and+Jun+Zhao%2E+2016a%2E+Leveraging+framenet+to+improve+automatic+event+detection%2E+In+Proceedings+of+the+54th+Annual+Meeting+of+the+Association+for+Computationa) [our card](https://md.research.elamur.ai/free/en/papers/liu-2016-leveraging-framenet-improve-automatic-event/)
17. Shulin Liu, Kang Liu, Shizhu He, and Jun Zhao. 2016b. A probabilistic soft logic based approach to exploiting latent and global information in event classification. In Proceedings of the thirtieth AAAI Conference on Artificail Intelligence. pages 2993–2999. [http://www.aaai.org/ocs/index.php/AAAI](http://www.aaai.org/ocs/index.php/AAAI) /AA… [↗](http://www.aaai.org/ocs/index.php/AAAI)
18. Haitao Mi, Zhiguo Wang, and Abe Ittycheriah. 2016. Supervised attentions for neural machine translation. arXiv preprint arXiv:1608.00112 https://arxiv.org/abs/1608.00112. [↗](https://arxiv.org/abs/1608.00112)
19. Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013. Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781 https://arxiv.org/abs/1301.3781. [↗](https://arxiv.org/abs/1301.3781)
20. Huu Thien Nguyen, Kyunghyun Cho, and Ralph Grishman. 2016. Joint event extraction via recurrent neural networks. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Association for Computational Linguistic… [find](https://scholar.google.com/scholar?q=Huu+Thien+Nguyen%2C+Kyunghyun+Cho%2C+and+Ralph+Grishman%2E+2016%2E+Joint+event+extraction+via+recurrent+neural+networks%2E+In+Proceedings+of+the+2016+Conference+of+the+North+American+Chapter+of+the+Association+) [our card](https://md.research.elamur.ai/free/en/papers/nguyen-2016-joint-event-extraction-recurrent-neural/)
21. Sebastian Riedel, Hong-Woo Chun, Toshihisa Tak-Proceed-agi, and Jun’ichi Tsujii. 2009. ings of the bionlp 2009 workshop companion volume for shared task. Association for Computational Linguistics, pages 41–49. [http://aclweb.org/anthology/W09-1406](http://aclweb.org/anthology/W09-1406). [↗](http://aclweb.org/anthology/W09-1406)
22. Deepak Venugopal, Chen Chen, Vibhav Gogate, and Vincent Ng. 2014. Relieving the computational bottleneck: Joint inference for event extraction with high-dimensional features. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Association for Computatio… [find](https://scholar.google.com/scholar?q=Deepak+Venugopal%2C+Chen+Chen%2C+Vibhav+Gogate%2C+and+Vincent+Ng%2E+2014%2E+Relieving+the+computational+bottleneck%3A+Joint+inference+for+event+extraction+with+high%2Ddimensional+features%2E+In+Proceedings+of+the+201)
23. Matthew D Zeiler. 2012. Adadelta: An adaptive learning rate method. arXiv preprint arXiv:1212.5701 https://arxiv.org/abs/1212.5701. [↗](https://arxiv.org/abs/1212.5701) [our card](https://md.research.elamur.ai/free/en/papers/zeiler-2012-adadelta-adaptive-learning-rate-method/)
24. Zhu Zhu, Shoushan Li, Guodong Zhou, and Rui Xia. 2014. Bilingual event extraction: a case study on trigger type determination. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). Association for Computational Linguistics, pages 842–847… [find](https://scholar.google.com/scholar?q=Zhu+Zhu%2C+Shoushan+Li%2C+Guodong+Zhou%2C+and+Rui+Xia%2E+2014%2E+Bilingual+event+extraction%3A+a+case+study+on+trigger+type+determination%2E+In+Proceedings+of+the+52nd+Annual+Meeting+of+the+Association+for+Computat)
25. Huu Thien Nguyen and Ralph Grishman. 2015. Event detection and domain adaptation with convolutional neural networks. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 2: Short P… [find](https://scholar.google.com/scholar?q=Huu+Thien+Nguyen+and+Ralph+Grishman%2E+2015%2E+Event+detection+and+domain+adaptation+with+convolutional+neural+networks%2E+In+Proceedings+of+the+53rd+Annual+Meeting+of+the+Association+for+Computational+Ling) [our card](https://md.research.elamur.ai/free/en/papers/nguyen-2015-event-detection-domain-adaptation-convolutional/)
26. Huu Thien Nguyen and Ralph Grishman. 2016. Modeling skip-grams for event detection with convolutional neural networks. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, pages 886–891. [http://aclweb.org/anthology/D16-…](http://aclweb.org/anthology/D16-…) [↗](http://aclweb.org/anthology/D16-…)
27. Siddharth Patwardhan and Ellen Riloff. 2009. A unified model of phrasal and sentential evidence for information extraction. In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, pages 151–160. [http://aclweb.org/anthology…](http://aclweb.org/anthology…) [↗](http://aclweb.org/anthology…)
28. Hoifung Poon and Lucy Vanderwende. 2010. Joint inference for knowledge extraction from [find](https://scholar.google.com/scholar?q=Hoifung+Poon+and+Lucy+Vanderwende%2E+2010%2E+Joint+inference+for+knowledge+extraction+from)