Scientific discovery has always been cumulative, from making new observations and extracting insights within existing data to connecting ideas across disciplines. AI offers the possibility of accelerating this process at every level, while the resulting knowledge can fuel increasingly capable AI systems. This symbiotic relationship creates the potential for knowledge flywheels, where advances in AI and scientific understanding continually reinforce one another. In this talk, I will highlight recent advances that are making these flywheels possible: combining generative models with scientific knowledge to solve inverse problems, developing multimodal foundation models that learn and connect representations across diverse forms of data, and building AI agents that distill, curate, and accumulate knowledge from experience. Together, these advances point toward a future in which scientific progress compounds through increasingly powerful cycles of learning and discovery.
Bio: Yisong Yue is a Professor of Computing and Mathematical Sciences at the Caltech. His research studies how to scale learning across the full cycle from data to discovery. He develops AI systems that learn efficiently from diverse data, improve through inference and accumulated experience, and connect models to scientific workflows and experiments. His prior work includes contributions to online learning and sequential decision-making, with applications spanning autonomous driving, safety-critical robotics, Mars rover navigation, spinal cord stimulation, information retrieval, and sports analytics. His work has been recognized with multiple paper awards and nominations across machine learning and its applications. He also currently serves on the ICLR board.
To request accommodations for a disability please contact Emily Lawrence, emilyl@cs.princeton.edu, at least one week prior to the event.