[13 minutes] Universal Approximation with Deep Narrow Networks
The classical universal approximation theorem (dating back to ~1999) is a foundational result, giving (part of) an answer to the question "why do neural networks work?" This is a brief introduction to our recent paper which proves some natural "dual" theorems on universal approximation, and in particular highlights a difference between deep and shallow neural networks.
Paper: https://arxiv.org/abs/1905.08539
Short version of this video: • [50 seconds] Universal Approximation with ...
This presentation was prepared as part of the acceptance of our paper to the Conference on Learning Theory 2020.