Abstract: To address the issues of inadequate extraction of intra-modal information and inaccurate cross-modal aspect semantic alignment in existing multimodal aspect-based sentiment analysis, a ...