Dependency Parsing with Backtracking using Deep Reinforcement Learning

TL;DR

This paper introduces a reinforcement learning approach for dependency parsing that allows backtracking to correct errors, improving accuracy by exploring alternative solutions in NLP tasks.

Contribution

It proposes a novel backtracking method using deep reinforcement learning to mitigate error propagation in dependency parsing and POS tagging.

Findings

Backtracking improves parsing accuracy.

Reinforcement learning effectively guides backtracking decisions.

Method reduces error propagation in NLP parsing tasks.

Abstract

Greedy algorithms for NLP such as transition based parsing are prone to error propagation. One way to overcome this problem is to allow the algorithm to backtrack and explore an alternative solution in cases where new evidence contradicts the solution explored so far. In order to implement such a behavior, we use reinforcement learning and let the algorithm backtrack in cases where such an action gets a better reward than continuing to explore the current solution. We test this idea on both POS tagging and dependency parsing and show that backtracking is an effective means to fight against error propagation.