Deep Learning Games
Full Text: nips16a.pdf
We investigate a reduction of supervised learning to game playing that reveals new connections and learning methods. For convex one-layer problems, we demonstrate an equivalence between global minimizers of the training problem and Nash equilibria in a simple game. We then show how the game can be extended to general acyclic neural networks with differentiable convex gates, establishing a bijection between the Nash equilibria and critical (or KKT) points of the deep learning problem. Based on these connections we investigate alternative learning methods, and find that regret matching can achieve competitive training performance while producing sparser models than current deep learning strategies.
Citation
D. Schuurmans, M. Zinkevich. "Deep Learning Games". Neural Information Processing Systems (NIPS), (ed: D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, R. Garnett), pp 1678-1686, December 2016.| Keywords: | |
| Category: | In Conference |
| Web Links: | NeurIPS |
BibTeX
@incollection{Schuurmans+Zinkevich:NIPS16,
author = {Dale Schuurmans and Martin Zinkevich},
title = {Deep Learning Games},
Editor = {D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, R. Garnett},
Pages = {1678-1686},
booktitle = {Neural Information Processing Systems (NIPS)},
year = 2016,
}Last Updated: February 14, 2020Submitted by Sabina P
