Ryan Yi

Ryan is interested in the nature of cognition in artificial systems–particularly what it means for a system to truly reason, abstract, or possess understanding. He aims to explore the full range of ways in which a system might be said to perform cognition, whether by mimicking natural intelligence or through other mechanisms. His focus draws on broader interests in neuroscience, linguistics, and the philosophy of mind.

At SFI, Ryan is working with Melanie Mitchell to investigate multimodal reasoning in AI systems. This includes evaluating existing models on established benchmarks for visual reasoning, designing new and more targeted benchmarks, and developing a novel neurosymbolic system that integrates prior knowledge and structured visual routines.

Before joining SFI, Ryan received a B.A. in Computer Science and Statistics from Harvard University.