学术讲座

讲座时间：2024年2月23日 10：00—11：30

腾讯会议：425 491 392

邀请人：宋睿华 高瓴人工智能学院长聘副教授

主讲人：陈师哲 法国国家信息与自动化研究所研究员

报告题目: Connecting 3D vision and robotics

报告摘要: Training robots to assist humans in daily tasks has been a longstanding dream in robotics. To achieve this goal, robots need to interact with humans, understand objects and scenes, and crucially, execute actions in the 3D world. While current advancements in large language models (LLMs) and large multimodal models (LMMs) have demonstrated remarkable success, their primary focus on text and 2D images leaves a critical gap in enabling robots to perceive and act in within 3D space. In this talk, I will discuss approaches to connect 3D visions and robotics in perception, reasoning, and action. First, I will introduce how to extend LLMs for situated 3D reasoning to answer complicated questions in 3D environments. Next, I will talk about using 3D perceptions to enhance language-guided robotic manipulation and present a recent work of 3D representation learning for robotics. Finally, I will address the reconstruction of 3D human-object interactions, which facilitates the potential for leveraging abundant online videos to learn dexterious actions for robots.

嘉宾简介: Shizhe Chen is a research scientist at WILLOW project-team in Inria Paris. She received her bachelor’s and PhD degrees at Renmin University of China in 2015 and 2020 respectively, supervised by Prof. Qin Jin. She then spent two wonderful post-doctoral years at Inria Paris collaborating with Dr. Ivan Laptev and Dr. Cordelia Schmid. Shizhe’s primary interests lie in embodied AI, vision and language, and multimodal deep learning. She has published over 30 peer-reviewed papers in leading conferences in computer vision, machine learning and robotics such as CVPR, ICCV, ECCV, NeurIPS, ACM Multimedia, CoRL and IROS. She also served as area chairs in CVPR, ICCV, ECCV, NeurIPS, ACM Multimedia and ICLR.