Sequential Cross-Document Coreference Resolution

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 4659–4671 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 4659 Sequential Cross-Document Coreference Resolution Emily Allaway ∗1 Shuai Wang 2 1 Department of Computer Science, Columbia University, New York, NY 2 Amazon AI eallaway@cs.columbia.edu {wshui, ballemig}@amazon.com Miguel Ballesteros 2 Abstract Relating entities and events in text is a key component of natural language understanding. Cross-document coreference resolution, in par- ticular, is important for the growing interest in multi-document analysis tasks. In this work we propose a new model that extends the effi- cient sequential prediction paradigm for coref- erence resolution to cross-document settings and achieves competitive results for both en- tity and event coreference while providing strong evidence of the efficacy of both se- quential models and higher-order inference in cross-document settings. Our model incremen- tally composes mentions into cluster represen- tations and predicts links between a mention and the already constructed clusters, approxi- mating a higher-order model. In addition, we conduct extensive ablation studies that provide new insights into the importance of various in- puts and representation types in coreference. 1 Introduction Relating entities and events in text is a key com- ponent of natural language understanding. For ex- ample, whether two news articles describing hurri- canes are referring to the same hurricane event. A crucial component of answering such questions is reasoning about groups entities and events across multiple documents. The goal of coreference resolution is to com- pute these clusterings of entities or events from extracted spans of text. While within-document coreference has been studied extensively (e.g., Lee et al. (2017, 2018)), there has been relatively less work on the cross-document task. However, grow- ing interest in multi-document applications, such as summarization (e.g., Liu and Lapata (2019); Fabbri et al. (2019)) and reading comprehension (e.g., Yan et al. (2019); Welbl et al. (2018)), highlights the importance of developing efficient and accurate ∗ Work done during internship at Amazon AI cross-document coreference models to minimize error-propagation in complex reasoning tasks. In this work we focus on cross-document coref- erence (CDCR), which implicitly requires within- document coreference (WDCR), and propose a new model that improves both coreference per- formance and computational complexity. Recent advances in within-document entity coreference resolution have shown that sequential prediction (i.e., making coreference predictions from left to right in a text1) achieves strong performance (Lee et al., 2017) with lower computational costs. This paradigm is also well suited to real-world streaming settings, where new documents are received every day, since it can easily find corefering events and entities with already processed documents, while most non-sequential models would require a full re- run. In this work, we show how this technique can first be extended to cross-document entity coref- erence and then adapted to cross-document event coreference. Our method is also able to take advantage of the history of previously made coreference decisions, approximating a higher-order model (i.e., operating on mentions as well as structures with mentions). Specifically, for every mention, a coreference …

Content truncated to 500 words