Understanding Human Behavior for Robotic Manipulation

Date: 2023/01/04 - 2023/01/04

Academic Seminar: Understanding Human Behavior for Robotic Manipulation

Speaker: Haoshu Fang, doctoral student of Shanghai Jiao Tong University

Time: 11:00 a.m. Jan. 4, 2022 (Beijing Time)

Abstract

Our society will be benefited a lot if we can build robots that are smart enough to fulfill many of the things humans do, and eventually work in a household environment seamlessly. How can robots acquire diverse skills efficiently to achieve this goal? A promising path is to teach a robot new skills like teaching a kid by demonstration, instead of re-programming the robot. This requires a breakthrough in several fields of AI, especially human behavior understanding and robotic manipulation. In this talk, Hao-Shu Fang will introduce his previous research in these two areas, including his representative work AlphaPose and AnyGrasp. In addition, Hao-Shu will share his thinking about how to combine these two topics in his future research.

Biography

Fang Hao-Shu is a fourth-year doctoral student of Shanghai Jiao Tong University, supervised by Prof. Cewu Lu. He is selected into the Zhiyuan Honorable Doctor Program and Wu Wenjun's class. His main research direction is human behavior understanding and robot manipulation. His research vision is to give robots dexterous manipulation skills to replace humans' tedious work in the industrial, agricultural and service industry. His representative projects are AlphaPose and AnyGrasp. During his undergraduate and PhD period, Fang Hao-Shu published 21 papers at TPAMI, RAL, Nature, CVPR, ICCV, ECCV, AAAI, ICRA, etc. The total citation is over 3000 times. He also serves as reviewer for TPAMI, TRO, IJCV, TMM, TOC, CVPR, ICCV, ECCV, ICLR, etc. Fang Hao-Shu was awarded the 2019 CCF-CV Rising-Star Award, 2019 Baidu Fellowship, 2020 Microsoft Fellowship, and 2021 ByteDance Fellowship. Personal homepage: https://fang-haoshu.github.io