Researchers have developed CF-Net, a deep multimodal network designed to recognize ambivalence and hesitancy in videos. This network utilizes frozen SigLIP2, HuBERT, and DistilBERT backbones to process visual, audio, and transcript data. CF-Net incorporates a ConflictFusion module to compute cross-modal incongruence and speaker normalization to reduce identity leakage. The model achieved strong performance on the BAH dataset, reaching a Macro F1 score of 0.7364 on the private challenge test set. AI
IMPACT This research advances multimodal AI capabilities in understanding subtle human expressions like ambivalence and hesitancy.
RANK_REASON The cluster contains an academic paper detailing a new model and its performance on a specific challenge.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
