CUET_SYNTHETICA@DravidianLangTech 2026: Multilingual Transformer Based Hope Speech Detection for Coarse and Fine-Grained Classification in Tulu

2026-06-25 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Natural Language Processing · Depth: Expert, medium

Summary

The CUET_SYNTHETICA team at DravidianLangTech@ACL 2026 presented a study on multilingual transformer-based hope speech detection for code-mixed Tulu, a low-resource language. Their work addressed two subtasks: coarse-grained classification into Encouraging, Discouraging, Uninvolved, and Blended categories (Task 1), and fine-grained classification into Optimistic, Realistic, Inspiring, Fading, and Hopelessness (Task 2). They fine-tuned XLM-RoBERTa-base, MuRIL, and mBERT multilingual transformer encoders. For Task 1, a soft-voting ensemble of all three models achieved a macro F1 of 0.58, securing 1st place. In Task 2, XLM-RoBERTa-base alone outperformed the other models, reaching a macro F1 of 0.42 and also securing 1st place. This research highlights effective strategies for NLP in under-resourced linguistic contexts.

Key takeaway

For NLP Engineers developing solutions for low-resource or code-mixed languages like Tulu, you should consider fine-tuning multilingual transformer encoders. An ensemble approach, specifically soft-voting with models like XLM-RoBERTa-base, MuRIL, and mBERT, significantly improves coarse-grained classification performance. For more nuanced, fine-grained tasks, prioritize XLM-RoBERTa-base as it demonstrated superior individual performance. This strategy can help you achieve competitive results in under-explored linguistic domains.

Key insights

Multilingual transformers effectively detect hope speech in low-resource, code-mixed Tulu across coarse and fine-grained categories.

Principles

• Ensemble models can boost classification performance.
• XLM-RoBERTa-base excels in fine-grained text classification.
• Multilingual models are beneficial for low-resource languages.

Method

Fine-tuning XLM-RoBERTa-base, MuRIL, and mBERT multilingual transformer encoders, followed by a three-way soft-voting ensemble for coarse-grained classification.

In practice

• Apply soft-voting ensembles for multi-label classification tasks.
• Prioritize XLM-RoBERTa-base for nuanced text classification.
• Explore multilingual models for code-mixed language data.

Topics

• Hope Speech Detection
• Tulu Language
• Multilingual Transformers
• Code-Mixed NLP
• XLM-RoBERTa-base
• Low-Resource Languages

Best for: Research Scientist, AI Scientist, NLP Engineer, Machine Learning Engineer

Related on AIssential

• mdok-style at SemEval-2026 Task 9: Finetuning LLMs for Multilingual Polarization Detection — Finetuning mid-sized LLMs with QLoRA and augmented data improves multilingual online polarization detection.
• TeamHerald@CHIPSAL 2026: Hate Speech Detection and Sentiment Analysis of Nepali Memes using Transformer-based Architectures and Ensemble Learning — Ensemble strategies for meme analysis perform differently for binary vs. multi-class tasks, requiring task-specific aggregation.
• An End-to-End Hybrid Framework for Rumour Detection in Low-Resources Algerian Dialect — A hybrid framework combining transformer embeddings with classical classifiers enables effective rumour detection in low-resource Algerian dialect.
• Indigenous Letters to Brazil: Multi-Label Classification — Multi-label classification of indigenous letters presents challenges due to corpus imbalance and semantic overlap.
• AFRILANGTUTOR: Advancing Language Tutoring and Culture Education in Low-Resource Languages with Large Language Models — Dictionary-based synthetic data generation and combined SFT+DPO significantly enhance LLM tutoring for low-resource African languages.
• Cross-lingual Relation Extraction with Large Language Models: Zero-Shot, Few-Shot, and Fine-Tuned Evaluation on Romanian — QLoRA fine-tuning significantly boosts cross-lingual relation extraction performance for low-resource languages, outperforming prompt-only LLMs.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.