This talk will present the latest research advancements of Yanwei Fu’s research group from the School of Big Data at Fudan University, focusing on the domains of multimodal models, visually guided robotic grasping, and fMRI neural decoding. The presentation will cover, but is not restricted to, topics such as few-shot learning, object-centered self-supervised learning, research pertaining to image inpainting, editing, motion, and other manipulations in multimodal computational photography, efficient alignment between visual and language modules in large multimodal models, efficient sparse compression techniques for Transformer models, as well as the utilization of large models and MVS 3D modeling in applications like visually guided robotic arm grasping.

Dr. Yanwei Fu, a full professor at the School of Data Science of Fudan University, obtained his PhD from Queen Mary University of London in 2014. Subsequently, he served as a post-doctoral researcher at Disney Research in Pittsburgh (co-located with CMU) from January 2015 to July 2016. In acknowledgment of his research achievements, Dr. Fu was honored with the Australia DECRA Award in 2016 and the 1000 Young Talent Scholar in 2018. Additionally, he holds the position of Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning. Throughout his productive and outstanding career, Dr. Fu has published more than 150 papers (with over 15K Google scholar citations), including 15 first-/corresponding-authorship papers in IEEE TPAMI. His research encompasses a wide array of topics, such as one-shot learning for images and videos, 3D reconstruction, vision-guided robotics, image generation/editing, and fMRI neural decoding.