Weighed Domain-Invariant Representation Learning for Cross-domain Sentiment Analysis

TL;DR

This paper introduces a weighted domain-invariant representation learning (WDIRL) framework that improves cross-domain sentiment analysis by addressing label distribution shifts, enhancing the transferability of existing models.

Contribution

It proposes a novel WDIRL framework that modifies existing DIRL models to better handle label distribution changes across domains.

Findings

WDIRL outperforms standard DIRL in cross-domain sentiment tasks.

Empirical results confirm the effectiveness of WDIRL in real-world scenarios.

Abstract

Cross-domain sentiment analysis is currently a hot topic in the research and engineering areas. One of the most popular frameworks in this field is the domain-invariant representation learning (DIRL) paradigm, which aims to learn a distribution-invariant feature representation across domains. However, in this work, we find out that applying DIRL may harm domain adaptation when the label distribution $\rm{P}(\rm{Y})$ changes across domains. To address this problem, we propose a modification to DIRL, obtaining a novel weighted domain-invariant representation learning (WDIRL) framework. We show that it is easy to transfer existing SOTA DIRL models to WDIRL. Empirical studies on extensive cross-domain sentiment analysis tasks verified our statements and showed the effectiveness of our proposed solution.