ICTS:32496

Turing lecture: Dynamical phenomena in nonlinear learning

(2025). Turing lecture: Dynamical phenomena in nonlinear learning. SciVideos. https://scivideos.org/icts-tifr/32496

Turing lecture: Dynamical phenomena in nonlinear learning. SciVideos, Aug. 13, 2025, https://scivideos.org/icts-tifr/32496 

          @misc{ scivideos_ICTS:32496,
            doi = {},
            url = {https://scivideos.org/icts-tifr/32496},
            author = {},
            keywords = {},
            language = {en},
            title = {Turing lecture: Dynamical phenomena in nonlinear learning},
            publisher = {},
            year = {2025},
            month = {aug},
            note = {ICTS:32496 see, \url{https://scivideos.org/icts-tifr/32496}}
          }
Andrea Montanari
Talk numberICTS:32496
Source RepositoryICTS-TIFR

Abstract

The success of modern AI models defies classical theoretical wisdom. Classical theory recommended the use of convex optimization, and yet AI models learn by optimizing highly non-convex function. Classical theory prescribed to control model complexity and yet AI models are very complex, so complex that they often memorize the training data. Classical wisdom recommends a careful and interpretable choice of model architecture, and yet modern architectures rarely offer a parsimonious representation of a target distribution class.

The discovery that learning can take place in completely unexpected scenario poses beautiful conceptual challenges. I will try to survey recent work towards addressing them.