Improving Disentangled Text Representation Learning with Information Theoretical Guidance

Publication Date: 7/5/2020

Event: ACL 2020

Reference: pp.7530-7541, 2020

Authors: Pengyu Cheng, NEC Laboratories America, Inc., Duke University; Martin Renqiang Min, NEC Laboratories America, Inc.; Dinghan Shen, NEC Laboratories America, Inc., Microsoft Dynamics 365 AI; Christopher Malon, NEC Laboratories America, Inc.; Yizhe Zhang, NEC Laboratories America, Inc.; Yitong Li, NEC Laboratories America, Inc., Microsoft Research; Lawrence Carin, Duke University

Abstract: Learning disentangled representations of natural language is essential for many NLP tasks, e.g., conditional text generation, style transfer, personalized dialogue systems, etc. Similar problems have been studied extensively for other forms of data, such as images and videos. However, the discrete nature of natural language makes the disentangling of textual representations more challenging (e.g., the manipulation over the data space cannot be easily achieved). Inspired by information theory, we propose a novel method that effectively manifests disentangled representations of text, without any supervision on semantics. A new mutual information upper bound is derived and leveraged to measure dependence between style and content. By minimizing this upper bound, the proposed method induces style and content embeddings into two independent low-dimensional spaces. Experiments on both conditional text generation and text-style transfer demonstrate the high quality of our disentangled representation in terms of content and style preservation.

Publication Link: