Continual Learning in Artificial Intelligence (AI)

This course explores one of the most important open questions in artificial intelligence: how can intelligent systems continue to learn after they are deployed? While modern AI models can retrieve information and reason over vast amounts of context, they remain fundamentally static, unable to continuously update their knowledge and skills through experience. This course examines the foundations of continual learning, including memory, adaptation, catastrophic forgetting, in-context learning, external memory systems, test-time learning, meta-learning, and self-improving agents. Drawing from machine learning, cognitive science, and systems research, we will investigate how future AI systems may move beyond static models toward lifelong learning, and what that means for intelligence, scientific discovery, and the design of human-centered technologies.