Abstract:

Multi-agent interacting systems are prevalent in the world, from purely physical systems to complicated social dynamics systems. The interactions between entities / components can give rise to very complex behavior patterns at the level of both individuals and the whole system. In many real-world multi-agent interacting systems (e.g., traffic participants, mobile robots, sports players), it is likely that only the trajectories of individual entities can be observed without knowledge of the underlying relations or interaction patterns. There may also be uncertainty and even multiple potential modalities in the behaviors of agents. Therefore, it is challenging to model the dynamics of the multi-agent system. In this talk, I will discuss our recent work on learning-based relational reasoning for multi-agent interacting systems with an emphasis on graph-based approaches. I will also talk about its application in various downstream tasks, such as motion prediction, multi-target tracking and decision making of intelligent agents. Finally, I will give a brief summary about the talk and discuss potential future research directions.

BIO:

Jiachen Li is currently a final-year Ph.D. candidate supervised by Prof. Masayoshi Tomizuka at University of California, Berkeley. He is working as a graduate student researcher at Mechanical Systems Control (MSC) Laboratory and Berkeley DeepDrive (BDD). His current research interest lies at the intersection of machine learning, graph neural networks, computer vision approaches and their applications to scene understanding, relational reasoning, motion prediction and decision making for intelligent systems (e.g. autonomous vehicles, mobile robots). He has served as a co-organizer and program committee member of multiple workshops on autonomous driving and a reviewer of journals / conferences on machine learning, computer vision and robotics.

Homepage:  https://jiachenli94.github.io/

Sponsored in part by:   Facebook Reality Labs Pittsburgh