Disentangled Representation Learning for Non-Parallel Text Style Transfer

Understand

This paper tackles the problem of disentangling the latent variables of style and content in language models.

• We propose a simple yet effective approach, which incorporates auxiliary multi-task and adversarial objectives, for label prediction and bag-of-words prediction, respectively.
• We show, both qualitatively and quantitatively, that the style and content are indeed disentangled in the latent space.
• This disentangled latent representation learning method is applied to style transfer on non-parallel corpora.