KIT-TIP-NLP at MultiPride: Continual Learning with Multilingual Foundation Model

In three linesMulti-stage framework for detecting reclaimed slurs in multilingual social media (English, Spanish, Italian). Uses XLM-RoBERTa with GPT-4o-mini back-translation data augmentation (×3 corpus), dynamic undersampling, and language-specific threshold optimization. Achieves 2-5% absolute F1 improvement without model retraining.

Summary generated by Claude — human-verified