SJM_MINDS@DravidianLangTech@ACL2026: Machine Learning Approaches for Hope Speech Detection in Code-Mixed Tulu

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

Summary

The SJM_MINDS system, submitted to the DravidianLangTech@ACL 2026 shared task, addresses hope speech detection in code-mixed Tulu. This system participated in two tasks: Coarse-Grained Hope Tone Classification (Task 1) and Fine-Grained Hope Type Classification (Task 2). Researchers experimented with classical Machine Learning approaches, specifically Logistic Regression (LR) and Linear Support Vector Classifier (LinearSVC), trained on Term Frequency and Inverse Document Frequency (TF-IDF) of word ngrams (n=1, 2). LinearSVC achieved a macro F1-score of 0.51 in Task 1, securing 4th rank, while the LR model obtained a macro F1-score of 0.37 in Task 2, ranking 5th. These results highlight the continued effectiveness of traditional ML methods for low-resource, code-mixed language scenarios.

Key takeaway

For NLP Engineers developing text classification systems for low-resource or code-mixed languages, you should consider classical Machine Learning approaches. Your team can achieve competitive performance by employing models like Logistic Regression and LinearSVC, especially when combined with TF-IDF word ngrams. This strategy offers a robust and often simpler alternative to complex deep learning models, proving effective for tasks like hope speech detection.

Key insights

Traditional ML models effectively detect hope speech in low-resource, code-mixed Tulu.

Principles

• Classical ML models remain effective for low-resource NLP.
• TF-IDF with n-grams is a viable feature extraction method.

Method

Train Logistic Regression and LinearSVC models using TF-IDF word ngrams (n=1, 2) for coarse and fine-grained hope speech classification in code-mixed text.

In practice

• Apply LinearSVC for coarse-grained text classification.
• Use Logistic Regression for fine-grained classification.
• Consider TF-IDF n-grams for low-resource languages.

Topics

• Hope Speech Detection
• Code-Mixed Languages
• Tulu Language
• Machine Learning
• TF-IDF
• Text Classification

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

Related on AIssential

• Resume Text Classifier using Classical NLP — Classical NLP and Logistic Regression can effectively classify resume text into job categories.
• Deep Learning Based Amharic Chatbot for FAQs in Universities — A deep learning chatbot effectively answers Amharic university FAQs, achieving 91.55% accuracy.
• Low-resource Language Discrimination Towards Chinese Dialects with Transfer learning and Data Augmentation — Transfer learning and data augmentation effectively overcome scarce annotation resources for low-resource Chinese dialect discrimination.
• Indigenous Letters to Brazil: Multi-Label Classification — Multi-label classification of indigenous letters presents challenges due to corpus imbalance and semantic overlap.
• Token Classification — Token classification is foundational for NLP, enabling machines to understand text at a granular, word-by-word level.
• From Text to Numbers: Demystifying Text Preprocessing in NLP — NLP text preprocessing, using methods like Bag of Words and TF-IDF, translates human language into numerical data for machine learning models.

Open in AIssential →

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