A Critical Analysis of Biased Parsers in Unsupervised Parsing

Abstract

A series of recent papers has used a parsing algorithm due to Shen et al. (2018) to recover phrase-structure trees based on proxies for "syntactic depth." These proxy depths are obtained from the representations learned by recurrent language models augmented with mechanisms that encourage the (unsupervised) discovery of hierarchical structure latent in natural language sentences. Using the same parser, we show that proxies derived from a conventional LSTM language model produce trees comparably well to the specialized architectures used in previous work. However, we also provide a detailed analysis of the parsing algorithm, showing (1) that it is incomplete---that is, it can recover only a fraction of possible trees---and (2) that it has a marked bias for right-branching structures which results in inflated performance in right-branching languages like English. Our analysis shows that evaluating with biased parsing algorithms can inflate the apparent structural competence of language models.

References in corpus (2)

• Adam: A Method for Stochastic Optimization
• Visually Grounded Neural Syntax Acquisition

Cited by in corpus (6)

• Are Pre-trained Language Models Aware of Phrases? Simple but Strong Baselines for Grammar Induction
• Perturbed Masking: Parameter-free Probing for Analyzing and Interpreting BERT
• What is Learned in Visually Grounded Neural Syntax Acquisition
• Exploiting Syntactic Structure for Better Language Modeling: A Syntactic Distance Approach
• On the Branching Bias of Syntax Extracted from Pre-trained Language Models
• Latent Tree Learning with Ordered Neurons: What Parses Does It Produce?