Decomposing Textual Information For Style Transfer
Ivan P. Yamshchikov, Viacheslav Shibaev, Aleksander Nagaev, J\"urgen, Jost, Alexey Tikhonov

TL;DR
This paper investigates how to effectively decompose textual information into different aspects using style transfer, proposing empirical assessment methods validated by state-of-the-art models, linking decomposition quality to translation performance.
Contribution
It introduces empirical methods for evaluating information decomposition in style transfer, enhancing understanding of how decomposed representations impact transfer quality.
Findings
Higher decomposition quality correlates with better BLEU scores.
Validated methods across multiple style transfer models.
Improved understanding of information representation in text style transfer.
Abstract
This paper focuses on latent representations that could effectively decompose different aspects of textual information. Using a framework of style transfer for texts, we propose several empirical methods to assess information decomposition quality. We validate these methods with several state-of-the-art textual style transfer methods. Higher quality of information decomposition corresponds to higher performance in terms of bilingual evaluation understudy (BLEU) between output and human-written reformulations.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
