Research Interests

It is believed that stimuli may activate “neurons that perceive different modalities” in the human brain, and the multimodal nature of its processing has inspired researchers to design multimodal-based tasks and solutions. In the context of modern deep learning, with the help of big models and big data, especially in generative scenarios, this special interest group not only focuses on representation, understanding, reasoning and generation in text or images, but also aims to better discover Multimodal phenomena, defining multimodal tasks, designing new models for multimodal learning.