DuNST: Dual Noisy Self Training for Semi-Supervised Controllable Text Generation

TL;DR

DuNST introduces a novel semi-supervised approach for controllable text generation that enhances exploration of the text space by jointly modeling generation and classification with noise-perturbed pseudo data, improving control accuracy.

Contribution

The paper proposes DuNST, a dual noisy self-training method that jointly models text generation and classification using a shared Variational AutoEncoder, with noise to improve exploration and generalization in controllable text generation.

Findings

Significantly improves control accuracy in generation tasks.

Maintains fluency and diversity comparable to strong baselines.

Theoretically guarantees enhanced exploration of the text space.

Abstract

Self-training (ST) has prospered again in language understanding by augmenting the fine-tuning of pre-trained language models when labeled data is insufficient. However, it remains challenging to incorporate ST into attribute-controllable language generation. Augmented by only self-generated pseudo text, generation models over-emphasize exploitation of the previously learned space, suffering from a constrained generalization boundary. We revisit ST and propose a novel method, DuNST to alleviate this problem. DuNST jointly models text generation and classification with a shared Variational AutoEncoder and corrupts the generated pseudo text by two kinds of flexible noise to disturb the space. In this way, our model could construct and utilize both pseudo text from given labels and pseudo labels from available unlabeled text, which are gradually refined during the ST process. We theoretically demonstrate that DuNST can be regarded as enhancing exploration towards the potential real text space, providing a guarantee of improved performance. Experiments on three controllable generation tasks show that DuNST could significantly boost control accuracy while maintaining comparable generation fluency and diversity against several strong baselines.