• Cross-modal fusion and alignment methods（跨模态融合与对齐方法)
• Affective computing and emotion recognition  (情感计算与情绪识别)
• Social signal processing and behavioral understanding  (社会信号处理与人类行为理解)
• Physiological signal analysis and multimodal biosensing  (生理信号分析与多模态生物传感)
• Advances in generative foundation models  (多模态基础模型前沿探索)
• Trustworthy multimodal AI: Evaluation, Interpretability, and Hallucination Mitigation  (多模态评估、可解释性与幻觉抑制)
• Embodied perception and interaction  (具身感知与交互)
• Datasets, benchmarks and real-world deployment of multimodal systems  (多模态智能系统的数据集、基准测试与实际部署)

Multi-modal perception and affective computing are transforming intelligent systems by enabling machines to understand human behavior, emotions, and intentions through diverse data sources such as vision, speech, language, physiological signals, and contextual information. This workshop provides a forum for researchers and practitioners to present recent advances in multi-modal representation learning, cross-modal fusion, emotion recognition, social signal processing, foundation models, and human-centered AI. It also addresses key challenges including robustness, generalization, explainability, fairness, privacy, and real-world deployment. By fostering interdisciplinary collaboration and discussion, the workshop aims to promote innovative research and practical applications of affect-aware, multi-modal intelligent systems across healthcare, education, robotics, and human–computer interaction.