Entailment Graph Learning with Textual Entailment and Soft Transitivity

Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics Volume 1: Long Papers, pages 5899 - 5910 May 22-27, 2022 c⃝2022 Association for Computational Linguistics Entailment Graph Learning with Textual Entailment and Soft Transitivity Zhibin Chen123 Yansong Feng13∗ Dongyan Zhao123 1Wangxuan Institute of Computer Technology, Peking University, China 2Center for Data Science, Peking University, China 3The MOE Key Laboratory of Computational Linguistics, Peking University, China {czb-peking, fengyansong, zhaody}@pku.edu.cn Abstract Typed entailment graphs try to learn the en- tailment relations between predicates from text and model them as edges between predicate nodes. The construction of entailment graphs usually suffers from severe sparsity and unrelia- bility of distributional similarity. We propose a two-stage method, Entailment Graph with Tex- tual Entailment and Transitivity (EGT2). EGT2 learns local entailment relations by recogniz- ing possible textual entailment between tem- plate sentences formed by typed CCG-parsed predicates. Based on the generated local graph, EGT2 then uses three novel soft transitivity con- straints to consider the logical transitivity in en- tailment structures. Experiments on benchmark datasets show that EGT2 can well model the transitivity in entailment graph to alleviate the sparsity issue, and lead to significant improve- ment over current state-of-the-art methods1. 1 Introduction Entailment, as an important relation in natural language processing (NLP), is critical to seman- tic understanding and natural language inference (NLI). Entailment relation has been widely applied in different NLP tasks such as Question Answer- ing (Pathak et al., 2021; Khot et al., 2018), Machine Translation (Padó et al., 2009) and Knowledge Graph Completion (Yoshikawa et al., 2019). When coming across a question that "Which medicine cures the infection?", one can recognize the infor- mation "Griseofulvin is preferred for the infection," in the corpus and appropriately write down the an- swer with the knowledge that "is preferred for" entails "cures" when their arguments are medicines and diseases, although the surface form of predi- cate "cures" does not exactly appear in the corpus. There are many ways to present one question, and it is impossible to handle them without understanding ∗Corresponding author. 1Our code is available at https://github.com/ ZacharyChenpk/EGT2. [medicine] is prefer- red for [disease] [medicine] cures [disease] [medicine] is effective for [disease] [medicine] is related to [disease] [medicine] causes [disease] t1=medicine t2=disease [medicine] is prefer- red for [disease] [medicine] cures [disease] [medicine] is effective for [disease] [medicine] is related to [disease] [medicine] causes [disease] t1=medicine t2=disease Figure 1: A simple example of entailment graph with types medicine and disease. The dashed line repre- sents a missing entailment recovered by considering the transitivity constraint (red) based on the two premise entailment between three boldfaced predicates. the entailment relations behind the predicates. Pre- vious works on analyzing entailment mainly focus on Recognizing Textual Entailment (RTE) between pairs of sentences, and many recent attempts have achieved quite promising performance in detecting entailment relations using transformer-based lan- guage models (He et al., 2020; Raffel et al., 2020; Schmitt and Schütze, 2021b). By modeling typed predicates as nodes and entailment relations as directed edges, the …

Content truncated to 500 words