Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning

TL;DR

This paper introduces a reinforcement learning approach combined with unlikelihood training to fine-tune language models, effectively reducing repetition and improving the quality of generated text.

Contribution

It proposes a novel method that integrates policy gradient reinforcement learning with unlikelihood training for better neural text generation.

Findings

Reduces repetition in generated text

Maintains language model quality

Outperforms baseline methods in control metrics

Abstract

Likelihood training and maximization-based decoding result in dull and repetitive generated texts even when using powerful language models (Holtzman et al., 2019). Adding a loss function for regularization was shown to improve text generation output by helping avoid unwanted properties, such as contradiction or repetition (Li at al., 2020). In this work, we propose fine-tuning a language model by using policy gradient reinforcement learning, directly optimizing for better generation. We apply this approach to minimizing repetition in generated text, and show that, when combined with unlikelihood training (Welleck et al., 2020), our method further reduces repetition without impacting the language model quality. We also evaluate other methods for improving generation at training and decoding time, and compare them using various metrics aimed at control for better text generation output.