TY - JOUR

T1 - Quick Estimate of Information Decomposition for Text Style Transfer

AU - Shibaev, Viacheslav

AU - Olbrich, Eckehard

AU - Jost, Jürgen

AU - Yamshchikov, Ivan P.

PY - 2023

Y1 - 2023

N2 - A growing number of papers on style transfer for texts rely on information decomposition. The performance of the resulting systems is usually assessed empirically in terms of the output quality or requires laborious experiments. This paper suggests a straightforward information theoretical framework to assess the quality of information decomposition for latent representations in the context of style transfer. Experimenting with several state-of-the-art models, we demonstrate that such estimates could be used as a fast and straightforward health check for the models instead of more laborious empirical experiments.

AB - A growing number of papers on style transfer for texts rely on information decomposition. The performance of the resulting systems is usually assessed empirically in terms of the output quality or requires laborious experiments. This paper suggests a straightforward information theoretical framework to assess the quality of information decomposition for latent representations in the context of style transfer. Experimenting with several state-of-the-art models, we demonstrate that such estimates could be used as a fast and straightforward health check for the models instead of more laborious empirical experiments.

UR - http://www.scopus.com/inward/record.url?partnerID=8YFLogxK&scp=85148940853

UR - https://gateway.webofknowledge.com/gateway/Gateway.cgi?GWVersion=2&SrcAuth=tsmetrics&SrcApp=tsm_test&DestApp=WOS_CPL&DestLinkType=FullRecord&KeyUT=000944926300001

U2 - 10.3390/e25020322

DO - 10.3390/e25020322

M3 - Article

VL - 25

JO - Entropy

JF - Entropy

SN - 1099-4300

IS - 2

M1 - 322

ER -