Deep learning is built on the foundational guarantee that gradient descent on an objective function converges to local minima. Unfortunately, this guarantee fails in settings, such as generative adversarial nets, that exhibit multiple interacting losses. The behavior of gradient-based methods in games is not well understood and is becoming increasingly important as adversarial and multi-objective architectures proliferate. In this paper, we develop new tools to understand and control the dynamics in n-player differentiable games. The key result is to decompose the game Jacobian into two components. The first, symmetric component, is related to potential games, which reduce to gradient descent on an implicit function. The second, antisymmetric component, relates to Hamiltonian games, a new class of games that obey a conservation law akin to conservation laws in classical mechanical systems. The decomposition motivates Symplectic Gradient Adjustment (SGA), a new algorithm for finding stable fixed points in differentiable games. Basic experiments show SGA is competitive with recently proposed algorithms for finding stable fixed points in GANs – while at the same time being applicable to, and having guarantees in, much more general cases.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Differentiable Game Mechanics


    Beteiligte:
    Letcher, A (Autor:in) / Balduzzi, D (Autor:in) / Racaniere, S (Autor:in) / Martens, J (Autor:in) / Foerster, J (Autor:in) / Tuyls, K (Autor:in) / Graepel, T (Autor:in)

    Erscheinungsdatum :

    2019-01-20


    Anmerkungen:

    Journal of Machine Learning Research , 20 (84) pp. 1-40. (2019)


    Medientyp :

    Aufsatz (Zeitschrift)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Klassifikation :

    DDC:    629



    The Mechanics of n-Player Differentiable Games

    Balduzzi, D / Racaniere, S / Martens, J et al. | BASE | 2018

    Freier Zugriff

    The ChainQueen differentiable physics engine

    Hu, Yuanming,S. M.Massachusetts Institute of Technology. | BASE | 2019

    Freier Zugriff

    Differentiable Rendering for Synthetic Aperture Radar Imagery

    Wilmanski, Michael C. / Tamir, Jonathan I. | IEEE | 2024


    Stable Opponent Shaping in Differentiable Games

    Letcher, A / Foerster, J / Balduzzi, D et al. | BASE | 2019

    Freier Zugriff

    Differentiable Rendering for Pose Estimation in Proximity Operations

    Bhaskara, Ramchander / Eapen, Roshan T. / Majji, Manoranjan | AIAA | 2023