Speaker
Anindita Maiti
(Perimeter Institute for Theoretical Physics)
Description
Abstract: In this short talk, I will provide the overview to an interplay between quantum field theory and theory of AI. In this foundational crosstalk, (a) field theories are simulated with guaranteed accuracy via dual neural networks (the backbones of AI) with no learning dynamics; (b) field theories additionally generate novel AI models; (c) AI learning dynamics corresponds to information theoretic RG flows; and lastly, (d) Wilsonian RG explains unforeseen learning behavior in large AI models. Altogether, this integrated paradigm marks the onset of a new era where similar synergies would enrich both AI theory and fundamental physics.
External references
- 24100123
- 4151144c-8fc6-4747-a5ff-74529fb1a32d