Unsupervised Evaluation Metrics and Learning Criteria for Non-Parallel Textual Transfer

TL;DR

This paper introduces new unsupervised evaluation metrics and training strategies for non-parallel textual transfer, improving the assessment of semantic preservation, fluency, and transfer accuracy.

Contribution

It proposes a comprehensive set of metrics and loss functions for better evaluation and training of non-parallel text transfer models, addressing current limitations.

Findings

Metrics correlate well with human judgments

Significant improvements over baseline methods

Effective in both sentiment and literature datasets

Abstract

We consider the problem of automatically generating textual paraphrases with modified attributes or properties, focusing on the setting without parallel data (Hu et al., 2017; Shen et al., 2017). This setting poses challenges for evaluation. We show that the metric of post-transfer classification accuracy is insufficient on its own, and propose additional metrics based on semantic preservation and fluency as well as a way to combine them into a single overall score. We contribute new loss functions and training strategies to address the different metrics. Semantic preservation is addressed by adding a cyclic consistency loss and a loss based on paraphrase pairs, while fluency is improved by integrating losses based on style-specific language models. We experiment with a Yelp sentiment dataset and a new literature dataset that we propose, using multiple models that extend prior work (Shen et al., 2017). We demonstrate that our metrics correlate well with human judgments, at both the sentence-level and system-level. Automatic and manual evaluation also show large improvements over the baseline method of Shen et al. (2017). We hope that our proposed metrics can speed up system development for new textual transfer tasks while also encouraging the community to address our three complementary aspects of transfer quality.