To Beam Or Not To Beam: That is a Question of Cooperation for Language GANs

TL;DR

This paper introduces SelfGAN, a cooperative framework for training language GANs that improves stability and performance by enabling discriminator and generator to produce cooperative outputs, leading to state-of-the-art results in summarization and question generation.

Contribution

It proposes a novel cooperative training approach for language GANs, enhancing stability and performance over traditional methods.

Findings

SelfGAN outperforms Teacher Forcing in experiments.

SelfGAN achieves state-of-the-art results on Summarization.

SelfGAN improves training stability for language GANs.

Abstract

Due to the discrete nature of words, language GANs require to be optimized from rewards provided by discriminator networks, via reinforcement learning methods. This is a much harder setting than for continuous tasks, which enjoy gradient flows from discriminators to generators, usually leading to dramatic learning instabilities. However, we claim that this can be solved by making discriminator and generator networks cooperate to produce output sequences during training. These cooperative outputs, inherently built to obtain higher discrimination scores, not only provide denser rewards for training, but also form a more compact artificial set for discriminator training, hence improving its accuracy and stability. In this paper, we show that our SelfGAN framework, built on this cooperative principle, outperforms Teacher Forcing and obtains state-of-the-art results on two challenging tasks, Summarization and Question Generation.