Keywords and Instances: A Hierarchical Contrastive Learning Framework Unifying Hybrid Granularities for Text Generation

TL;DR

This paper introduces a hierarchical contrastive learning framework that unifies multiple granularities of semantic meaning, especially keywords, to improve text generation tasks like paraphrasing and storytelling.

Contribution

It proposes a novel hierarchical contrastive learning mechanism that incorporates keyword-level and instance-level contrasts, addressing the contrast vanishing problem and enhancing generation quality.

Findings

Outperforms baselines on paraphrasing, dialogue generation, and storytelling.

Effectively models keyword contributions in contrastive learning.

Addresses contrast vanishing issue with inter-contrast mechanism.

Abstract

Contrastive learning has achieved impressive success in generation tasks to militate the "exposure bias" problem and discriminatively exploit the different quality of references. Existing works mostly focus on contrastive learning on the instance-level without discriminating the contribution of each word, while keywords are the gist of the text and dominant the constrained mapping relationships. Hence, in this work, we propose a hierarchical contrastive learning mechanism, which can unify hybrid granularities semantic meaning in the input text. Concretely, we first propose a keyword graph via contrastive correlations of positive-negative pairs to iteratively polish the keyword representations. Then, we construct intra-contrasts within instance-level and keyword-level, where we assume words are sampled nodes from a sentence distribution. Finally, to bridge the gap between independent contrast levels and tackle the common contrast vanishing problem, we propose an inter-contrast mechanism that measures the discrepancy between contrastive keyword nodes respectively to the instance distribution. Experiments demonstrate that our model outperforms competitive baselines on paraphrasing, dialogue generation, and storytelling tasks.