Not All Pretraining are Created Equal: Threshold Tuning and Class Weighting for Imbalanced Polarization Tasks in Low-Resource Settings
It addresses polarization classification in low-resource settings, but is incremental as it applies existing methods to a specific shared task.
The paper tackles polarization detection in social media text for English and Swahili, achieving up to 0.8032 macro-F1 on binary detection and up to 0.556 on multi-label tasks using Transformer models with techniques like class weighting and threshold tuning.
This paper describes my submission to the Polarization Shared Task at SemEval-2025, which addresses polarization detection and classification in social media text. I develop Transformer-based systems for English and Swahili across three subtasks: binary polarization detection, multi-label target type classification, and multi-label manifestation identification. The approach leverages multilingual and African language-specialized models (mDeBERTa-v3-base, SwahBERT, AfriBERTa-large), class-weighted loss functions, iterative stratified data splitting, and per-label threshold tuning to handle severe class imbalance. The best configuration, mDeBERTa-v3-base, achieves 0.8032 macro-F1 on validation for binary detection, with competitive performance on multi-label tasks (up to 0.556 macro-F1). Error analysis reveals persistent challenges with implicit polarization, code-switching, and distinguishing heated political discourse from genuine polarization.