Aspects of Similarity in Natural and Artificial Minds

Abstract: Appeals to similarity have been a key part of how cognitive scientists have studied mental representations, following the pioneering work of Shepard and Tversky. In this talk, I will re-examine this tradition using the modern lens of cognitive science, leveraging AI, Bayesian modeling, and novel human-in-the-loop paradigms that emulate statistical algorithms. I will argue that these tools not only allow us to test classic results more rigorously, but also to explore new questions about similarity, alignment, and abstraction in humans and machines that were previously intractable.