Exploiting Argument Information to Improve Event Detection via Supervised Attention Mechanisms
Open accessThe DOI matches its Crossref record; 28 of the 29 references it cites were found in a registry or an index.
Abstract
This paper tackles the task of event detection (ED), which involves identifying and categorizing events.We argue that arguments provide significant clues to this task, but they are either completely ignored or exploited in an indirect manner in existing detection approaches.In this work, we propose to exploit argument information explicitly for ED via supervised attention mechanisms.In specific, we systematically investigate the proposed model under the supervision of different attention strategies.Experimental results show that our approach advances state-ofthe-arts and achieves the best F 1 score on ACE 2005 dataset.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures
![Figure 1 The architecture of the proposed approach for event detection. In this figure, w is the candidate word, [ w 1 , ..., w n ] is the contextual words of w , and [ e 1 , ..., e n ] is the corresponding entity types of [ w 1 , ... , w n ].](https://figures.cochat.ai/paper-figures/6f24d8cdc0694dd4a06a3cab21a83367/exploiting-argument-information-to-improve-event-detection-via-supervised-figure-1.png)

Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“Supervised attention [76] aims to provide explicit supervision over the attention of models based on the fact that completely data-driven attention can overfit to some artifacts [73].”
From Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing (2023) · cites this paper as [76]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Efficient Estimation of Word Representations in Vector Space
Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1301.3781 · Cited by 26,568 · cited 2×
“Speci(cid:12)cally, we use the Skip-gram model (Mikolov et al., 2013) to learn word embed-dings on the NYT corpus 3 .”
MethodConvolutional Neural Networks for Sentence Classification
Kim, Y. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1181 · Cited by 14,831 · cited 1×
“Regularization is implemented by a dropout (Kim, 2014; Hinton et al., 2012) and L2 norm.”
BackgroundImproving neural networks by preventing co-adaptation of feature detectors
Hinton, G. E., Srivastava, N., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. R. (2012) · arXiv · doi:10.48550/arxiv.1207.0580 · Cited by 6,382 · cited 2×
“Regularization is implemented by a dropout (Kim, 2014; Hinton et al., 2012) and L2 norm.”
BackgroundADADELTA: An Adaptive Learning Rate Method
Zeiler, M. D. (2012) · arXiv · doi:10.48550/arxiv.1212.5701 · Cited by 5,322 · cited 1×
“We train the model by using a simple optimization technique called stochastic gradient descent (SGD) over shuffled mini-batches with the Adadelta rule (Zeiler, 2012).”
MethodHagan, M., Demuth, H. B., Beale, M. H., & Jes?s, O. D. (2016) · Institutional Repository of the National University Open and Distance UNAD (Universidad Nacional Abierta y a Distancia) · Cited by 5,133
A neural probabilistic language model
BengioYoshua, DucharmeRéjean, VincentPascal, & JanvinChristian (2003) · Journal of Machine Learning Research · doi:10.5555/944919.944966 · Cited by 2,644
Why Does Unsupervised Pre-training Help Deep Learning?
Erhan, D., Bengio, Y., Courville, A. C., Manzagol, P.-A., Vincent, P., & Bengio, S. (2010) · Journal of Machine Learning Research · Cited by 2,641
Baker, C. F., Fillmore, C. J., & Lowe, J. B. (1998) · Proceedings of the 36th annual meeting on Association for Computational Linguistics - · doi:10.3115/980845.980860 · Cited by 2,518 · cited 1×
“In order to further demonstrate the e(cid:11)ectiveness of our approach, we also use events from FrameNet (FN) (F. Baker et al., 1998) as extra training data, as the same as Liu et al. (2016a) to alleviate the data-sparseness problem for ED to augment the performance of the proposed approach.”
MethodEvent Extraction via Dynamic Multi-Pooling Convolutional Neural Networks
Chen, Y., Xu, L., Liu, K., Zeng, D., & Zhao, J. (2015) · Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.3115/v1/p15-1017 · Cited by 906 · cited 7×
“Representation-based methods typically represent candidate event mentions by embeddings and feed them into neural networks (Chen et al., 2015; Nguyen and Grishman, 2015; Liu et al., 2016a; Nguyen and Grishman, 2016).”
MethodInfluentialSelf-citationJoint Event Extraction via Recurrent Neural Networks
Nguyen, T. H., Cho, K., & Grishman, R. (2016) · Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/n16-1034 · Cited by 647 · cited 6×
“…learning based methods have been proposed, which tackled event detection and argu-ment extraction simultaneously (Riedel et al., 2009; Li et al., 2013; Venugopal et al., 2014; Nguyen et al., 2016), these approaches usually only make remarkable improvements to AE, but insigni(cid:12)cant to ED.”
BackgroundMethodInfluentialIncremental Joint Extraction of Entity Mentions and Relations
Li, Q., & Ji, H. (2014) · Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.3115/v1/p14-1038 · Cited by 535 · cited 1×
“Moreover, to obtain entity information, we use RPI Joint Information Extraction System7 (Li et al., 2013, 2014; Li and Ji, 2014) to label ACE entity mentions.”
MethodJoint Event Extraction via Structured Prediction with Global Features
Li, Q., Ji, H., & Huang, L. (2013) · Cited by 479
Event Detection and Domain Adaptation with Convolutional Neural Networks
Nguyen, T. H., & Grishman, R. (2015) · 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 Papers) · doi:10.3115/v1/p15-2060 · Cited by 408 · cited 3×
“Following existing work (Li et al., 2013; Chen et al., 2015; Nguyen and Grishman, 2015), we exploit the annotated entity information in our ED system.”
MethodRefining Event Extraction through Cross-Document Inference
Ji, H., & Grishman, R. (2008) · Cited by 400
Using Document Level Cross-Event Inference to Improve Event Extraction
Liao, S., & Grishman, R. (2010) · Cited by 293
Using Cross-Entity Inference to Improve Event Extraction
Hong, Y., Zhang, J., Ma, B., Yao, J., Zhou, G., & Zhu, Q. (2011) · Cited by 239
Supervised Attentions for Neural Machine Translation
Mi, H., Wang, Z., & Ittycheriah, A. (2016) · Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d16-1249 · Cited by 131 · cited 1×
“Inspired by Mi et al. (2016), we use a gaussian distribution g ( (cid:1) ) to model the attention distribution of words around arguments.”
MethodModeling Skip-Grams for Event Detection with Convolutional Neural Networks
Nguyen, T. H., & Grishman, R. (2016) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/d16-1085 · Cited by 127 · cited 3×
“Unfortunately, most existing methods performed event detection individually, where the annotated arguments in training set are totally ignored (Ji and Grishman, 2008; Gupta and Ji, 2009; Hong et al., 2011; Chen et al., 2015; Nguyen and Grishman, 2015; Liu et al., 2016a,b; Nguyen and Grishman, 2016).”
MethodLeveraging FrameNet to Improve Automatic Event Detection
Liu, S., Chen, Y., He, S., Liu, K., & Zhao, J. (2016) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/p16-1201 · Cited by 126 · cited 14×
“We believe the reason is that the training data of ACE 2005 corpus is insufficient to train a precise attention in an unsupervised manner, considering that data sparseness is an important issue of ED (Zhu et al., 2014; Liu et al., 2016a).”
BackgroundMethodInfluentialSelf-citationA unified model of phrasal and sentential evidence for information extraction
Patwardhan, S., & Riloff, E. (2009) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/1699510.1699530 · Cited by 122 · cited 1×
“, POS tags, dependency relations) into feature vectors (Ahn, 2006; Ji and Grishman, 2008; Patwardhan and Riloff, 2009; Gupta and Ji, 2009; Liao and Grishman, 2010; Hong et al., 2011; Liu et al., 2016b).”
BackgroundConstructing Information Networks Using One Single Model
Li, Q., Ji, H., HONG, Y., & Li, S. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1198 · Cited by 114 · cited 2×
“The second kind of approach, on the contrast, tackled event detection and argument extraction simultaneously, which is called joint approach (Riedel et al., 2009; Poon and Van-derwende, 2010; Li et al., 2013, 2014; Venu-gopal et al., 2014; Nguyen et al., 2016).”
BackgroundMethodJoint Inference for Knowledge Extraction from Biomedical Literature
Poon, H., & Vanderwende, L. (2010) · Cited by 109
Predicting unknown time arguments based on cross-event propagation
Gupta, P. K., & Ji, H. (2009) · doi:10.3115/1667583.1667697 · Cited by 76 · cited 2×
“Unfortunately, most existing methods performed event detection individually, where the annotated arguments in training set are totally ignored (Ji and Grishman, 2008; Gupta and Ji, 2009; Hong et al., 2011; Chen et al., 2015; Nguyen and Grishman, 2015; Liu et al., 2016a,b; Nguyen and Grishman, 2016).”
BackgroundMethodLiu, S., Liu, K., He, S., & Zhao, J. (2016) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v30i1.10375 · Cited by 70 · cited 14×
“Processing of Events from FN Liu et al. (2016a) detected events from FrameNet based on the observation that frames in FN are analogous to events in ACE (lexical unit of a frame $ trigger of an even-t, frame elements of a frame $ arguments of an event).”
BackgroundMethodInfluentialSelf-citationVenugopal, D., Chen, C., Gogate, V., & Ng, V. (2014) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1090 · Cited by 67 · cited 3×
“…learning based methods have been proposed, which tackled event detection and argu-ment extraction simultaneously (Riedel et al., 2009; Li et al., 2013; Venugopal et al., 2014; Nguyen et al., 2016), these approaches usually only make remarkable improvements to AE, but insigni(cid:12)cant to ED.”
BackgroundMethodSupervised Attentions for Neural Machine TranslationPreprint
Mi, H., Wang, Z., & Ittycheriah, A. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1608.00112 · Cited by 29 · cited 1×
“Inspired by Mi et al. (2016), we use a gaussian distribution g ( (cid:1) ) to model the attention distribution of words around arguments.”
MethodProceedings of the Workshop on Annotating and Reasoning about Time and Events
Boguraev, B., Muñoz, R., & Pustejovsky, J. D. (2006) · Cited by 24
Bilingual Event Extraction: a Case Study on Trigger Type Determination
Zhu, Z., Li, S., Zhou, G., & Xia, R. (2014) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.3115/v1/p14-2136 · Cited by 22 · cited 2×
“We believe the reason is that the training data of ACE 2005 corpus is insufficient to train a precise attention in an unsupervised manner, considering that data sparseness is an important issue of ED (Zhu et al., 2014; Liu et al., 2016a).”
Background
Research works printed but not linked: 1
- [25]25. Sebastian Riedel, Hong-Woo Chun, Toshihisa Tak- agi, and Jun'ichi Tsujii. 2009. Proceed- ings of the bionlp 2009 workshop compan- ion volume for shared task. Association for Computational Linguistics, pages 41-49. http://aclweb.org/anthology/W09-1406.
Cited by
Liu, P., Yuan, W., Fu, J., Jiang, Z., Hayashi, H., & Neubig, G. (2023) · ACM Computing Surveys · doi:10.1145/3560815 · Cited by 4,361
“Supervised attention [76] aims to provide explicit supervision over the attention of models based on the fact that completely data-driven attention can overit to some artifacts [73].”
BackgroundA General Survey on Attention Mechanisms in Deep Learning
Brauwers, G., & Frasincar, F. (2023) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2021.3126456 · Cited by 695
Liu, P., Yuan, W., Fu, J., Jiang, Z., Hayashi, H., & Neubig, G. (2021) · arXiv · doi:10.48550/arxiv.2107.13586 · Cited by 4,352
“Supervised attention [76] aims to provide explicit supervision over the attention of models based on the fact that completely data-driven attention can overit to some artifacts [73].”
BackgroundJointly Multiple Events Extraction via Attention-based Graph Information Aggregation
Liu, X., Luo, Z., & Huang, H. (2018) · Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d18-1156 · Cited by 397
“• The positional embedding vector of w i : If w c is the current word, we encode the relative distance i − c from w i to w c as a real-valued vector by looking up the randomly initialized position embedding table (Nguyen et al., 2016; Liu et al., 2017; Nguyen and Gr-ishman, 2018).”
Graph Convolutional Networks With Argument-Aware Pooling for Event Detection
Nguyen, T., & Grishman, R. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.12039 · Cited by 383
“Syntactic dependency graphs represents sentences as directed trees with head-modifier dependency arcs between related words (Mcdonald and Pereira 2006; Koo, Carreras, and Collins 2008).”
Background
Metadata
- Authors
- Liu, S., Chen, Y., Liu, K., & Zhao, J.
- Affiliations
- Chinese Academy of Sciences (CN)
- Venue
- Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Published
- 2017
- Pages
- 1789-1798
- ISSN
- 0736-587X (ISSN-L)
- OpenAlex
- W2739918945
- Cited by
- 250
- References
- 29
- Citations by year
- 2026: 2 · 2025: 10 · 2024: 15 · 2023: 33 · 2022: 37 · 2021: 55 · 2020: 36 · 2019: 34 · 2018: 25 · 2017: 3
- FWCI
- 12.22
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Proceedings of the conference - Association for Computational Linguistics. Meeting (publishedVersion, cc-by)
- Funders
- National Natural Science Foundation of China (61533018)
- National Key Research and Development Program of China (2014CB340503)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Sentiment Analysis and Opinion Mining
- Keywords
- event detection, argument mining, event argument extraction, attentional strategies
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.