N-ary Constituent Tree Parsing with Recursive Semi-Markov Model

Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, pages 2631–2642 August 1–6, 2021. ©2021 Association for Computational Linguistics 2631 N-ary Constituent Tree Parsing with Recursive Semi-Markov Model Xin Xin∗, Jinlong Li School of Computer Science and Technology, Beijing Institute of Technology, Beijing, China BJ ER Center of HVLIPCCA, BIT, Beijing, China {xxin,jllee}@bit.edu.cn Zeqi Tan Zhejiang University, Hangzhou, China zqtan@zju.edu.cn Abstract In this paper, we study the task of graph-based constituent parsing in the setting that binariza- tion is not conducted as a pre-processing step, where a constituent tree may consist of nodes with more than two children. Previous graph- based methods on this setting typically gen- erate hidden nodes with the dummy label in- side the n-ary nodes, in order to transform the tree into a binary tree for prediction. The limitation is that the hidden nodes break the sibling relations of the n-ary node’s children. Consequently, the dependencies of such sib- ling constituents might not be accurately mod- eled and is being ignored. To solve this limi- tation, we propose a novel graph-based frame- work, which is called “recursive semi-Markov model”. The main idea is to utilize 1-order semi-Markov model to predict the immediate children sequence of a constituent candidate, which then recursively serves as a child can- didate of its parent. In this manner, the de- pendencies of sibling constituents can be de- scribed by 1-order transition features, which solves the above limitation. Through experi- ments, the proposed framework obtains the F1 of 95.92% and 92.50% on the datasets of PTB and CTB 5.1 respectively. Specially, the recur- sive semi-Markov model shows advantages in modeling nodes with more than two children, whose average F1 can be improved by 0.3-1.1 points in PTB and 2.3-6.8 points in CTB 5.1. 1 Introduction There are two settings for constituent parsing mod- els, including binary tree parsing and n-ary tree parsing. In the former, the original constituent tree with n-ary nodes is converted into a binary tree by language-specific rules. The model first predicts the binary tree, and then converts it back. In the latter, the model directly predicts the n-ary tree without the intermediate step of binarization. ∗Xin Xin is the corresponding author. IP NP VP QP ϕ （a） （b） （c） ϕ C C C C C C C C C C C C C C C C CLP NP QP VP CLP NP QP VP M VV NN M VV NN CD CD CLP C C C C C C C C C C C C C C NP VP CP IP NP NP CD VV M NN VV NN NN 利 润 profit 完 成 complete 企 业 enterprise 投 资 invest 个 体 individual 三 百 万 3 million 元 yuan 利 润 profit 完 成 complete 三 百 万 3 million 元 yuan 利 润 profit 完 成 complete 三 百 万 3 million 元 yuan Figure 1: An n-ary node and the …

Content truncated to 500 words