We present the winning approach to the TRAC 2024 Shared Task on Offline Harm Potential Identification (HarmPot-ID). The task focused on low-resource Indian languages and consisted of two sub-tasks: 1a) predicting the offline harm potential and 1b) detecting the most likely target(s) of the offline harm. We explored low-source domain specific, cross-lingual, and monolingual transformer models and submitted the aggregate predictions from the MuRIL and BERT models. Our approach achieved 0.74 micro-averaged F1-score for sub-task 1a and 0.96 for sub-task 1b, securing the 1st rank for both sub-tasks in the competition.

In the context of the proliferation of multimodal hate speech related to the Russia-Ukraine conflict, we introduce a unified multimodal fusion system for detecting hate speech and its targets in text-embedded images. Our approach leverages the Twitter-based RoBERTa and Swin Transformer V2 models to encode textual and visual modalities, and employs the Multilayer Perceptron (MLP) fusion mechanism for classification. Our system achieved macro F1 scores of 87.27% for hate speech detection and 80.05% for hate speech target detection in the Multimodal Hate Speech Event Detection Challenge 2024, securing the 1st rank in both subtasks. We open-source the trained models at https://huggingface.co/Yestin-Wang