Paper
The Platonic Representation Hypothesis
We argue that representations in AI models, particularly deep networks, are converging. First, we survey many 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, we demonstrate convergence across data modalities: as vision models and language models get larger, they measure distance between datapo · Authors: Minyoung Huh, Brian Cheung, Tongzhou Wang, Phillip Isola · Published 2024-05-13 · arXiv:2405.07987
- Published
- 2024-05-13
- Authors
- 4
- arXiv
- 2405.07987
- Category
- cs.LG
Abstract
arXiv:2405.07987
Minyoung Huh, Brian Cheung, Tongzhou Wang, Phillip Isola
We argue that representations in AI models, particularly deep networks, are converging. First, we survey many 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, we 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. We 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, we discuss the implications of these trends, their limitations, and counterexamples to our analysis.