Emotion Classification in Low and Moderate Resource Languages

TL;DR

This paper develops a cross-lingual emotion classification method trained on resource-rich languages and successfully transfers it to low and moderate-resource languages, demonstrating improved accuracy over baselines and creating new annotated datasets.

Contribution

It introduces two transfer learning approaches for emotion classification across languages and provides empirical evidence of their effectiveness on six diverse languages.

Findings

Direct transfer outperforms annotation projection.

Proposed methods outperform random baselines.

Annotated emotion datasets created for four languages.

Abstract

It is important to be able to analyze the emotional state of people around the globe. There are 7100+ active languages spoken around the world and building emotion classification for each language is labor intensive. Particularly for low-resource and endangered languages, building emotion classification can be quite challenging. We present a cross-lingual emotion classifier, where we train an emotion classifier with resource-rich languages (i.e. \textit{English} in our work) and transfer the learning to low and moderate resource languages. We compare and contrast two approaches of transfer learning from a high-resource language to a low or moderate-resource language. One approach projects the annotation from a high-resource language to low and moderate-resource language in parallel corpora and the other one uses direct transfer from high-resource language to the other languages. We show the efficacy of our approaches on 6 languages: Farsi, Arabic, Spanish, Ilocano, Odia, and Azerbaijani. Our results indicate that our approaches outperform random baselines and transfer emotions across languages successfully. For all languages, the direct cross-lingual transfer of emotion yields better results. We also create annotated emotion-labeled resources for four languages: Farsi, Azerbaijani, Ilocano and Odia.