Multi-task Learning for Cross-Lingual Sentiment Analysis

TL;DR

This paper introduces a multi-task, multilingual BERT-based model for cross-lingual sentiment analysis, enabling sentiment classification of Croatian news articles using Slovene data in zero-shot and few-shot settings.

Contribution

It proposes a simple multi-task learning approach with a trilingual BERT model for cross-lingual sentiment analysis, addressing low-resource language scenarios.

Findings

Effective sentiment classification in Croatian using Slovene data.

Multi-task setup improves cross-lingual transfer performance.

Zero-shot and few-shot scenarios demonstrate model robustness.

Abstract

This paper presents a cross-lingual sentiment analysis of news articles using zero-shot and few-shot learning. The study aims to classify the Croatian news articles with positive, negative, and neutral sentiments using the Slovene dataset. The system is based on a trilingual BERT-based model trained in three languages: English, Slovene, Croatian. The paper analyses different setups using datasets in two languages and proposes a simple multi-task model to perform sentiment classification. The evaluation is performed using the few-shot and zero-shot scenarios in single-task and multi-task experiments for Croatian and Slovene.