Speaker: Prof. Ding Zhao, Carnegie Mellon University

Time: 15:00 p.m., December 6, 2023, GMT+8

Venue: Room 204, Courtyard No.5, Jingyuan, PKU

Abstract: 

As AI becomes more integrated into new fields such as robotics and healthcare, it presents a dual spectrum of opportunities and risks. In this talk, I will introduce our efforts in deploying trustworthy intelligent autonomy at a large scale for vital civil usage such as self-driving cars, assistant robots, and autonomous surgery. During the deployment and transition, training data often exhibit significant imbalance, multi-modal complexity, and nonstationarity. I will initiate the discussion by analyzing 'long-tailed' problems with rare events and their connection to safety evaluation and safe reinforcement learning. I will then discuss how modeling multi-modal uncertainties as ‘tasks’ may enhance generalizability by learning across domains. To facilitate task delineation with high-dimensional inputs in vision and language, we have developed prompt-transformer-based structures for efficient adaptation and mitigation of catastrophic forgetting. In cases involving unknown-unknown tasks with severely limited data, we explore the potential of leveraging external knowledge from legislative sources, causal reasoning, and large language models. Lastly, we will expand intelligence development into the realm of system-level design space with meta physical robot morphologies, which may achieve generalizability and safety more effectively than relying solely on software solutions.

Source: Center on Frontiers of Computing Studies, PKU