A Span-level Bidirectional Network for Aspect Sentiment Triplet Extraction

Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 427 - 443 December 7-11, 2022 ©2022 Association for Computational Linguistics Sentence-Incremental Neural Coreference Resolution Matt Grenander Shay B. Cohen School of InformaticsMark Steedman University of Edinburgh matt.grenander@ed.ac.uk scohen@inf.ed.ac.uk, m.steedman@ed.ac.uk Abstract We propose a sentence-incremental neural coreference resolution system which incre- mentally builds clusters after marking mention boundaries in a shift-reduce method. The sys- tem is aimed at bridging two recent approaches at coreference resolution: (1) state-of-the-art non-incremental models that incur quadratic complexity in document length with high com- putational cost, and (2) memory network- based models which operate incrementally but do not generalize beyond pronouns. For com- parison, we simulate an incremental setting by constraining non-incremental systems to form partial coreference chains before observ- ing new sentences. In this setting, our system outperforms comparable state-of-the-art meth- ods by 2 F1 on OntoNotes and 6.8 F1 on the CODI-CRAC 2021 corpus. In a conventional coreference setup, our system achieves 76.3 F1 on OntoNotes and 45.5 F1 on CODI-CRAC 2021, which is comparable to state-of-the-art baselines. We also analyze variations of our system and show that the degree of incremen- tality in the encoder has a surprisingly large effect on the resulting performance.1 1 Introduction Coreference Resolution (CR) is a task in which a system detects and resolves linguistic expressions that refer to the same entity. It is typically per- formed in two steps: in mention detection, the model predicts which expressions are referential, and in mention clustering, the model computes each mention’s antecedent. Many recently pro- posed systems follow a mention-pair formulation from Lee et al. (2017), in which all possible spans are ranked and then scored against each other. In particular, methods that augment this approach with large, pre-trained language models achieve state-of-the-art results (Joshi et al., 2019, 2020). 1Code is available at: https://github.com/mgrenander/sentence-incremental-coref In 2004, on the Waterfront Promenade originally constructed for viewing only the scenery of Hong Kong Island and Victoria Har- bor, the Hong Kong Tourism Board also con- structed the Avenue of Stars, memorializing Hong Kong’s 100-year film history . Figure 1: An example from the OntoNotes dataset which highlights the need for incremental systems to identify spans rather than tokens as mentions. The men- tions cannot be resolved solely from the prefix ‘Hong Kong’, and the clustering decision should be delayed until the full mention is observed. Despite impressive performance, these methods are computationally demanding. For a text with n tokens, they will score up to O(n2) spans, fol- lowed by up to O(n4) span comparisons. They also process documents non-incrementally, requiring ac- cess to the entire document before processing can begin. These properties present challenges when insufficient computational resources are available, or when the task setup is incremental, such as in di- alogue (e.g. Khosla et al. 2021). From a cognitive perspective, these methods are also unappealing be- cause research on “garden-path” effects show that humans resolve referring expressions incrementally (Altmann and Steedman, 1988). These drawbacks have led to renewed …

Content truncated to 500 words