Yuyang Shen

A Cross-Attention-Based Framework for Multimodal Sentiment Analysis

Multimodal Sentiment Analysis (MSA) aims to predict human sentiment by jointly modeling complementary information from textual, acoustic, and visual modalities. However, effectively exploiting heterogeneous multimodal features remains challenging due to semantic inconsistency, temporal misalignment, and noisy modality-...