Transformer-Based Model for Multilingual Hope Speech Detection

TL;DR

This study evaluates transformer-based hope speech detection for English and German on the PolyHope-M benchmark using RoBERTa and XLM-RoBERTa. It shows strong English performance but comparatively weaker German results, highlighting language-specific challenges in morphology and dataset balance. The authors propose a language-aware preprocessing and model-tuning strategy and suggest exploring larger multilingual models and hybrid features to improve cross-language performance. The work emphasizes the potential and limitations of current large language models for multilingual hope speech tasks and guides future improvements for more robust multilingual hope speech detection.

Abstract

This paper describes a system that has been submitted to the "PolyHope-M" at RANLP2025. In this work various transformers have been implemented and evaluated for hope speech detection for English and Germany. RoBERTa has been implemented for English, while the multilingual model XLM-RoBERTa has been implemented for both English and German languages. The proposed system using RoBERTa reported a weighted f1-score of 0.818 and an accuracy of 81.8% for English. On the other hand, XLM-RoBERTa achieved a weighted f1-score of 0.786 and an accuracy of 78.5%. These results reflects the importance of improvement of pre-trained large language models and how these models enhancing the performance of different natural language processing tasks.

Figures (1)

• Figure 1: Overall Architecture of the Proposed Model