Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
102 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

TacoERE: Cluster-aware Compression for Event Relation Extraction (2405.06890v1)

Published 11 May 2024 in cs.CL and cs.AI

Abstract: Event relation extraction (ERE) is a critical and fundamental challenge for natural language processing. Existing work mainly focuses on directly modeling the entire document, which cannot effectively handle long-range dependencies and information redundancy. To address these issues, we propose a cluster-aware compression method for improving event relation extraction (TacoERE), which explores a compression-then-extraction paradigm. Specifically, we first introduce document clustering for modeling event dependencies. It splits the document into intra- and inter-clusters, where intra-clusters aim to enhance the relations within the same cluster, while inter-clusters attempt to model the related events at arbitrary distances. Secondly, we utilize cluster summarization to simplify and highlight important text content of clusters for mitigating information redundancy and event distance. We have conducted extensive experiments on both pre-trained LLMs, such as RoBERTa, and LLMs, such as ChatGPT and GPT-4, on three ERE datasets, i.e., MAVEN-ERE, EventStoryLine and HiEve. Experimental results demonstrate that TacoERE is an effective method for ERE.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (45)
  1. Docbert: Bert for document classification. arXiv preprint arXiv:1904.08398.
  2. Docbert: Bert for document classification. ArXiv, abs/1904.08398.
  3. Longformer: The long-document transformer. arXiv preprint arXiv:2004.05150.
  4. Knowledge-enriched event causality identification via latent structure induction networks. In Proceedings of ACL-IJCNLP, pages 4862–4872.
  5. Faithful to the original: Fact aware neural abstractive summarization. In Proceedings of AAAI, 32.
  6. ERGO: Event relational graph transformer for document-level event causality identification. In Proceedings of COLING, pages 2118–2128.
  7. Palm: Scaling language modeling with pathways. arXiv preprint arXiv:2204.02311.
  8. BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of NAACL, pages 4171–4186.
  9. GSum: A general framework for guided neural abstractive summarization. In Proceedings of NAACL, pages 4830–4842.
  10. Towards event-level causal relation identification. In Proceedings of SIGIR, page 1828–1833.
  11. Modeling document-level causal structures for event causal relation identification. In Proceedings of NAACL, pages 1808–1817.
  12. Making pre-trained language models better few-shot learners. In Proceedings of ACL-IJCNLP, pages 3816–3830.
  13. Deep feature-based text clustering and its explanation. IEEE Transactions on Knowledge and Data Engineering, 34(8):3669–3680.
  14. Frame semantic-enhanced sentence modeling for sentence-level extractive text summarization. In Proceedings of EMNLP, pages 4045–4052.
  15. Frame semantics guided network for abstractive sentence summarization. Knowledge-Based Systems, 221:106973.
  16. EventOA: An event ontology alignment benchmark based on FrameNet and Wikidata. In Findings of the ACL, pages 10038–10052.
  17. Joint event and temporal relation extraction with shared representations and structured prediction. In Proceedings of EMNLP-IJCNLP, pages 434–444.
  18. Marti A. Hearst and Christian Plaunt. 1993. Subtopic structuring for full-length document access. In Proceedings of the SIGIR, page 59–68.
  19. Question answering as global reasoning over semantic abstractions. In In Proceedings of AAAI.
  20. Chin-Yew Lin. 2004. ROUGE: A package for automatic evaluation of summaries. In Text Summarization Branches Out, pages 74–81.
  21. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.
  22. Selecting optimal context sentences for event-event relation extraction. In Proceedings of AAAI, 36(10):11058–11066.
  23. A well-composed text is half done! composition sampling for diverse conditional generation. In Proceedings of ACL, pages 1319–1339.
  24. Planning with learned entity prompts for abstractive summarization. In Proceedings of TACL, 9:1475–1492.
  25. Joint reasoning for temporal and causal relations. In Proceedings of ACL, pages 2278–2288.
  26. OpenAI. 2023. Gpt-4 technical report.
  27. Setrank: Learning a permutation-invariant ranking model for information retrieval. In Proceedings of SIGIR, pages 499–508.
  28. Ramakanth Pasunuru and Mohit Bansal. 2018. Multi-reward reinforced summarization with saliency and entailment. In Proceedings of the NAACL, pages 646–653.
  29. A deep reinforced model for abstractive summarization. In Proceedings of the ICLR.
  30. Enhancement of short text clustering by iterative classification. In Natural Language Processing and Information Systems, pages 105–117.
  31. Get to the point: Summarization with pointer-generator networks. In Proceedings of ACL, pages 1073–1083.
  32. Lamda: Language models for dialog applications. arXiv preprint arXiv:2201.08239.
  33. Minh Tran Phu and Thien Huu Nguyen. 2021. Graph convolutional networks for event causality identification with rich document-level structures. In Proceedings of NAACL, pages 3480–3490.
  34. Attention is all you need. In Proceedings of NeurIPS, 30.
  35. Joint constrained learning for event-event relation extraction. In Proceedings of the EMNLP, pages 696–706.
  36. BiSET: Bi-directional selective encoding with template for abstractive summarization. In Proceedings of ACL, pages 2153–2162.
  37. MAVEN-ERE: A unified large-scale dataset for event coreference, temporal, causal, and subevent relation extraction. In Proceedings of the EMNLP, pages 926–941.
  38. Ronald J. Williams. 1992. Simple statistical gradient-following algorithms for connectionist reinforcement learning. Mach. Learn., 8(3–4):229–256.
  39. Document-level relation extraction with sentences importance estimation and focusing. In Proceedings of NAACL, pages 2920–2929.
  40. Discriminative reasoning for document-level relation extraction. In Findings of ACL-IJCNLP, pages 1653–1663.
  41. Temporal common sense acquisition with minimal supervision. In Proceedings of ACL, pages 7579–7589.
  42. Document-level relation extraction with adaptive thresholding and localized context pooling. In Proceedings of AAAI, 35(16):14612–14620.
  43. The Storyline Annotation and Representation Scheme (StaR): A Proposal. In Proceedings of the 2nd Workshop on Computing News Storylines. PID https://github.com/tommasoc80/EventStoryLine.
  44. Goran Glavas and Jan Snajder and Marie-Francine Moens and Parisa KordJamshidi. 2014. HiEve: A Corpus for Extracting Event Hierarchies from News Stories. In Proceedings of LREC. PID http://takelab.fer.hr/hievents.rar.
  45. MAVEN-ERE: A Unified Large-scale Dataset for Event Coreference, Temporal, Causal, and Subevent Relation Extraction. In Proceedings of EMNLP. PID https://github.com/THU-KEG/MAVEN-ERE.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (7)
  1. Yong Guan (18 papers)
  2. Xiaozhi Wang (51 papers)
  3. Lei Hou (127 papers)
  4. Juanzi Li (144 papers)
  5. Jeff Pan (4 papers)
  6. Jiaoyan Chen (85 papers)
  7. Freddy Lecue (36 papers)
Citations (1)
View on Semantic Scholar
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets