Joint Coreference Resolution and Character Linking for Multiparty Conversation

Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics, pages 539–548 April 19 - 23, 2021. ©2021 Association for Computational Linguistics 539 Joint Coreference Resolution and Character Linking for Multiparty Conversation Jiaxin Bai1, Hongming Zhang1, Yangqiu Song1, and Kun Xu2 1CSE, HKUST 2 Tencent AI Lab {jbai, hzhangal, yqsong}@cse.ust.hk, kxkunxu@tencent.com Abstract Character linking, the task of linking men- tioned people in conversations to the real world, is crucial for understanding the conver- sations. For the efficiency of communication, humans often choose to use pronouns (e.g., “she”) or normal phrases (e.g., “that girl”) rather than named entities (e.g., “Rachel”) in the spoken language, which makes linking those mentions to real people a much more challenging than a regular entity linking task. To address this challenge, we propose to incor- porate the richer context from the coreference relations among different mentions to help the linking. On the other hand, considering that finding coreference clusters itself is not a triv- ial task and could benefit from the global char- acter information, we propose to jointly solve these two tasks. Specifically, we propose C2, the joint learning model of Coreference resolu- tion and Character linking. The experimental results demonstrate that C2 can significantly outperform previous works on both tasks. Fur- ther analyses are conducted to analyze the con- tribution of all modules in the proposed model and the effect of all hyper-parameters. 1 Introduction Understanding conversations has long been one of the ultimate goals of the natural language process- ing community, and a critical step towards that is grounding all mentioned people to the real world. If we can achieve that, we can leverage our knowl- edge about these people (e.g., things that happened to them before) to better understand the conversa- tion. On the other hand, we can also aggregate the conversation information back to our under- standing about these people, which can be used for understanding future conversations that involve the same people. To simulate the real conversations and investigate the possibility for models to ground mentioned people, the character linking task was 72% 16% 12% Pronouns Personal Nouns Named Entities Figure 1: The composition of the mentions in conver- sations for character grounding. Over 88% of the men- tions are not named entities, which brings exceptional challenges when linking those to character entities. proposed (Chen and Choi, 2016). Specifically, it uses the transcripts of TV shows (i.e., Friends) as the conversations and asks the models to ground all person mentions to characters. Even though the character linking task can be viewed as a special case of the entity linking task, it is more challenging than the ordinary entity linking task for various reasons. First, the ordinary entity linking task often aims at linking named entities to external knowledge bases such as Wikipedia, where rich information (e.g., definitions) are avail- able. However, for the character linking task, we do not have the support of such rich knowledge base and all we have are the names of …

Content truncated to 500 words