CSAIL Forum with Prof Phillip Isola: The Platonic Representation Hypothesis
The Platonic Representation Hypothesis
Speaker: Phillip Isola, Associate Professor, CSAIL
Tuesday 12:00-1:00 EDT, April 15, 2025
In person: Hewlett 32-G882 in the Stata Center, 32 Vassar Street
and live stream via Zoom: Registration required
Abstract: I will argue that representations in different deep nets are converging. First, I will survey examples of convergence in the literature: over time and across multiple domains, the ways by which different neural networks represent data are becoming more aligned. Next, I will demonstrate convergence across data modalities: as vision models and language models get larger, they measure distance between datapoints in a more and more alike way. I will hypothesize that this convergence is driving toward a shared statistical model of reality, akin to Plato's concept of an ideal reality. We term such a representation the platonic representation and discuss several possible selective pressures toward it. Finally, I'll discuss the implications of these trends, their limitations, and counterexamples to our analysis.