TY  - GEN
UR  - https://openreview.net/forum?id=B1xMEerYvB
PB  - ICLR
N2  - With the success of modern machine learning, it is becoming increasingly important to understand and control how learning algorithms interact. Unfortunately, negative results from game theory show there is little hope of understanding or controlling general n-player games. We therefore introduce smooth markets (SM-games), a class of n-player games with pairwise zero sum interactions. SM-games codify a common design pattern in machine learning that includes some GANs, adversarial training, and other recent algorithms. We show that SM-games are amenable to analysis and optimization using first-order methods.
ID  - discovery10109590
A1  - Balduzzi, D
A1  - Czarnecki, WM
A1  - Anthony, T
A1  - Gemp, IM
A1  - Hughes, E
A1  - Leibo, JZ
A1  - Piliouras, G
A1  - Graepel, T
KW  - game theory
KW  -  optimization
KW  -  gradient descent
KW  -  adversarial learning
SP  - 1
AV  - public
Y1  - 2020///
EP  - 18
TI  - Smooth markets: A basic mechanism for organizing gradient-based learners
N1  - This version is the author accepted manuscript. For information on re-use, please refer to the publisher's terms and conditions.
ER  -