Joint Multi-Decoder Framework with Hierarchical Pointer Network for Frame Semantic Parsing

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 2570–2578 August 1–6, 2021. ©2021 Association for Computational Linguistics 2570 Joint Multi-Decoder Framework with Hierarchical Pointer Network for Frame Semantic Parsing Xudong Chen1,2, Ce Zheng1, Baobao Chang1,3∗ 1The MOE Key Laboratory of Computational Linguistics, Peking University, China 2School of Software and Microelectronics, Peking University, China 3Peng Cheng Laboratory, Shenzhen, China {xdc,zce1112zslx,chbb}@pku.edu.cn Abstract Current researches on frame semantic parsing include three subtasks, namely frame identifi- cation, argument identification and role clas- sification. Most of previous systems process these subtasks independently and ignore their interactions. We introduce a novel architec- ture based on multi-decoder strategy to han- dle these subtasks together. The multi-decoder strategy strengthens the interactions. More- over, we design a hierarchical pointer network for argument identification which reduces the computational complexity. To our best knowl- edge, it’s the first practice to introduce the pointer network into frame semantic parsing. The experiments show improvement over state of the art models on FrameNet dataset. 1 Introduction Frame semantic parsing is a fundamental study in Natural Language Processing. It aims to parse sen- tences into frame-style semantic structures defined in FrameNet (Baker et al., 1998). An example of frame-style semantic structures is shown in Figure 1. The word write.v is a target that evokes the frame called Text creation. The phrases underlined with green lines are called ar- guments. Author, Text and Form are roles (also called frame elements) the arguments play in this frame. Hence the frame semantic parsing contains three subtasks, namely frame identification, argu- ment identification and role classification. For a sentence with a given target, the frame identifi- cation is to disambiguate the frame for the target based on its contextual information, the argument identification is to identify the boundaries of all the arguments, and the role classification is to assign a semantic role to each argument we have found. Early work (Hermann et al., 2014; FitzGerald et al., 2015; Hartmann et al., 2017) on frame seman- ∗Corresponding author Figure 1: A sentence annotated the arguments and roles of frame Text creation. The arrow marks indicate the order of arguments identification and roles classifica- tion. tic parsing adopts pipeline strategy. Their models apply independent models to handle different sub- tasks which ignore the interactions among subtasks. Moreover, the pipeline strategy usually causes error propagation problem. The accuracy of frame iden- tification can become the bottleneck of the overall performance. Later work (Yang and Mitchell, 2017; Peng et al., 2018) processes all the subtasks jointly by optimizing them together during training. Their joint models show improvement over pipeline mod- els, which demonstrates the benefit of joint training strategy. However, their systems don’t have spe- cific design to model the interactions among the subtasks. To strengthen the interactions of subtasks, we propose a joint framework based on three task- specific decoders. The interactions in our frame- work are mainly reflected in two aspects. On one hand, the representations of both the target and its frame that derived from frame identification de- …

Content truncated to 500 words