A Conditional Splitting Framework for Efficient Constituency Parsing

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 5795–5807 August 1–6, 2021. ©2021 Association for Computational Linguistics 5795 A Conditional Splitting Framework for Efficient Constituency Parsing Thanh-Tung Nguyen†¶, Xuan-Phi Nguyen†¶, Shafiq Joty¶§, Xiaoli Li†¶ ¶Nanyang Technological University §Salesforce Research Asia †Institute for Infocomm Research, A-STAR Singapore {ng0155ng@e.;nguyenxu002@e.;srjoty@}ntu.edu.sg xlli@i2r.a-star.edu.sg Abstract We introduce a generic seq2seq parsing frame- work that casts constituency parsing problems (syntactic and discourse parsing) into a series of conditional splitting decisions. Our pars- ing model estimates the conditional probabil- ity distribution of possible splitting points in a given text span and supports efficient top- down decoding, which is linear in number of nodes. The conditional splitting formulation together with efficient beam search inference facilitate structural consistency without rely- ing on expensive structured inference. Cru- cially, for discourse analysis we show that in our formulation, discourse segmentation can be framed as a special case of parsing which allows us to perform discourse parsing without requiring segmentation as a pre-requisite. Ex- periments show that our model achieves good results on the standard syntactic parsing tasks under settings with/without pre-trained repre- sentations and rivals state-of-the-art (SoTA) methods that are more computationally ex- pensive than ours. In discourse parsing, our method outperforms SoTA by a good margin. 1 Introduction A number of formalisms have been introduced to analyze natural language at different linguistic lev- els. This includes syntactic structures in the form of phrasal and dependency trees, semantic struc- tures in the form of meaning representations (Ba- narescu et al., 2013; Artzi et al., 2013), and dis- course structures with Rhetorical Structure Theory (RST) (Mann and Thompson, 1988) or Discourse- LTAG (Webber, 2004). Many of these formalisms have a constituency structure, where textual units (e.g., phrases, sentences) are organized into nested constituents. For example, Figure 1 shows exam- ples of a phrase structure tree and a sentence-level discourse tree (RST) that respectively represent how the phrases and clauses are hierarchically or- ganized into a constituency structure. Developing efficient and effective parsing solutions has always been a key focus in NLP. In this work, we consider both phrasal (syntactic) and discourse parsing. In recent years, neural end-to-end parsing meth- ods have outperformed traditional methods that use grammar, lexicon and hand-crafted features. These methods can be broadly categorized based on whether they employ a greedy transition-based, a globally optimized chart parsing or a greedy top- down algorithm. Transition-based parsers (Dyer et al., 2016; Cross and Huang, 2016; Liu and Zhang, 2017; Wang et al., 2017) generate trees auto-regressively as a form of shift-reduce deci- sions. Though computationally attractive, the local decisions made at each step may propagate errors to subsequent steps due to exposure bias (Bengio et al., 2015). Moreover, there may be mismatches in shift and reduce steps, resulting in invalid trees. Chart based methods, on the other hand, train neural scoring functions to model the tree structure globally (Durrett and Klein, 2015; Gaddy et al., 2018; Kitaev …

Content truncated to 500 words