Xuexin Wei - Learning and Inference with Prior Knowledge - IPAM at UCLA

Recorded 25 September 2024. Xuexin Wei of the University of Texas at Austin presents "Learning and Inference with Prior Knowledge" at IPAM's Analyzing High-dimensional Traces of Intelligent Behavior Workshop.
Abstract: Intelligent systems adapt to the environments through learning and appropriate usage of prior knowledge. In neuroscience, how the prior is implemented in the brain remains an open question. In artificial intelligence, developing computational algorithms that efficiently learn prior distribution over complex stimulus space, e.g., natural images, represents one active research topic. I will present several ongoing projects in my lab that seek to address these questions. These projects range from cognitive neuroscience (how the learning of prior knowledge influences human behavior) to Neuro-AI (how to leverage prior knowledge to improve the behavioral robustness of deep networks).
Learn more online at: www.ipam.ucla....