Smooth markets: A basic mechanism for organizing gradient-based learners
This provides a foundational mechanism for organizing gradient-based learners, addressing a key problem in machine learning theory for researchers and practitioners.
The paper tackles the challenge of understanding and controlling interactions among multiple gradient-based learning algorithms by introducing smooth markets (SM-games), a class of n-player games with pairwise zero-sum interactions that model common patterns like GANs and adversarial training, and shows they are amenable to analysis and optimization using first-order methods.
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.