Tackling Error Propagation through Reinforcement Learning: A Case of Greedy Dependency Parsing

TL;DR

This paper applies reinforcement learning to greedy dependency parsing to reduce error propagation, improving accuracy and robustness while maintaining efficiency, and provides evidence that it effectively decreases error propagation.

Contribution

It introduces reinforcement learning into greedy dependency parsing, demonstrating improved accuracy and reduced error propagation compared to traditional methods.

Findings

Reinforcement learning improves labeled and unlabeled dependency accuracy.

It reduces the occurrence of error propagation in greedy parsers.

The approach maintains the efficiency of the original parser.

Abstract

Error propagation is a common problem in NLP. Reinforcement learning explores erroneous states during training and can therefore be more robust when mistakes are made early in a process. In this paper, we apply reinforcement learning to greedy dependency parsing which is known to suffer from error propagation. Reinforcement learning improves accuracy of both labeled and unlabeled dependencies of the Stanford Neural Dependency Parser, a high performance greedy parser, while maintaining its efficiency. We investigate the portion of errors which are the result of error propagation and confirm that reinforcement learning reduces the occurrence of error propagation.