This seminar series probes some of the central questions in the physics of learning:  What properties of data make it learnable? How is learned information encoded in neural representations? Do neural networks exhibit universal properties? Can we construct a thermodynamic theory of learning? What determines how performance scales with model size and computational resources? 

This seminar series probes some of the central questions in the physics of learning:  What properties of data make it learnable? How is learned information encoded in neural representations? Do neural networks exhibit universal properties? Can we construct a thermodynamic theory of learning? What determines how performance scales with model size and computational resources? 

This seminar series probes some of the central questions in the physics of learning:  What properties of data make it learnable? How is learned information encoded in neural representations? Do neural networks exhibit universal properties? Can we construct a thermodynamic theory of learning? What determines how performance scales with model size and computational resources? 

This seminar series probes some of the central questions in the physics of learning:  What properties of data make it learnable? How is learned information encoded in neural representations? Do neural networks exhibit universal properties? Can we construct a thermodynamic theory of learning? What determines how performance scales with model size and computational resources?