资源分类
》 资源 》论文》CONVERGENCE BEHAVIOUR OF SOME GRADIENT-BASED METHODS ON BILINEAR ZERO -S UM GAMES

CONVERGENCE BEHAVIOUR OF SOME GRADIENT-BASED METHODS ON BILINEAR ZERO -S UM GAMES

2020-01-02 | |  54 |   42 |   0

Abstract

Min-max formulations have attracted great attention in the ML community due to the rise of deep generative models and adversarial methods, while understanding the dynamics of gradient algorithms for solving such formulations has remained a grand challenge. As a first step, we restrict to bilinear zero-sum games and give a systematic analysis of popular gradient updates, for both simultaneous and alternating versions. We provide exact conditions for their convergence and find the optimal parameter setup and convergence rates. In particular, our results offer formal evidence that alternating updates converge “better” than simultaneous ones.

上一篇：ON UNIVERSAL EQUIVARIANT SET NETWORKS

下一篇：NETWORK DECONVOLUTION

用户评价
全部评价

热门资源

  • The Variational S...

    Unlike traditional images which do not offer in...

  • Stratified Strate...

    In this paper we introduce Stratified Strategy ...

  • Learning to learn...

    The move from hand-designed features to learned...

  • A Mathematical Mo...

    Direct democracy, where each voter casts one vo...

  • Joint Pose and Ex...

    Facial expression recognition (FER) is a challe...