Latent Tree Learning with Differentiable Parsers: Shift-Reduce Parsing and Chart Parsing
Computation and Language
2020-01-16 v1
Abstract
Latent tree learning models represent sentences by composing their words according to an induced parse tree, all based on a downstream task. These models often outperform baselines which use (externally provided) syntax trees to drive the composition order. This work contributes (a) a new latent tree learning model based on shift-reduce parsing, with competitive downstream performance and non-trivial induced trees, and (b) an analysis of the trees learned by our shift-reduce model and by a chart-based model.
Cite
@article{arxiv.1806.00840,
title = {Latent Tree Learning with Differentiable Parsers: Shift-Reduce Parsing and Chart Parsing},
author = {Jean Maillard and Stephen Clark},
journal= {arXiv preprint arXiv:1806.00840},
year = {2020}
}
Comments
ACL 2018 workshop on Relevance of Linguistic Structure in Neural Architectures for NLP