Discourse Representation Structure Parsing for Chinese

TL;DR

This paper investigates Chinese semantic parsing without labeled data, introduces a new evaluation test suite, and analyzes parsing difficulties, especially due to adverbs, comparing different modeling approaches.

Contribution

It presents a pipeline for collecting Chinese meaning representation data, a specialized test suite for evaluation, and insights into parsing challenges and model performance.

Findings

Adverbs significantly impact Chinese semantic parsing difficulty.

Machine translation plus English parser performs slightly worse than direct Chinese models.

A new evaluation framework helps diagnose Chinese parsing challenges.

Abstract

Previous work has predominantly focused on monolingual English semantic parsing. We, instead, explore the feasibility of Chinese semantic parsing in the absence of labeled data for Chinese meaning representations. We describe the pipeline of automatically collecting the linearized Chinese meaning representation data for sequential-to sequential neural networks. We further propose a test suite designed explicitly for Chinese semantic parsing, which provides fine-grained evaluation for parsing performance, where we aim to study Chinese parsing difficulties. Our experimental results show that the difficulty of Chinese semantic parsing is mainly caused by adverbs. Realizing Chinese parsing through machine translation and an English parser yields slightly lower performance than training a model directly on Chinese data.