On the Properties of the Softmax Function with Application in Game Theory and Reinforcement Learning Article Swipe

PDF

Related Concepts

Softmax function Reinforcement learning Function (biology) Reinforcement Monotone polygon Lipschitz continuity Connection (principal bundle) Monotonic function Game theory Convex function Mathematics Computer science Mathematical optimization Applied mathematics Regular polygon Artificial intelligence Mathematical analysis Mathematical economics Engineering Artificial neural network Geometry Evolutionary biology Biology Structural engineering

Bolin Gao , Lacra Pavel ·

YOU? · · 2017 · Open Access · · DOI: https://doi.org/10.48550/arxiv.1704.00805 · OA: W2606101940

In this paper, we utilize results from convex analysis and monotone operator theory to derive additional properties of the softmax function that have not yet been covered in the existing literature. In particular, we show that the softmax function is the monotone gradient map of the log-sum-exp function. By exploiting this connection, we show that the inverse temperature parameter determines the Lipschitz and co-coercivity properties of the softmax function. We then demonstrate the usefulness of these properties through an application in game-theoretic reinforcement learning.