Smooth Games Optimization and Machine Learning Workshop

Room 512 ABEF, Friday Dec 7th, NeurIPS2018 , Montreal.

Videos of the Workshop
The videos of the workshop are available on our youtube channel.
The surveys recommended by the panel can be found Here .
We would like to thank Aaron Defazio, Alex Dimakis, David Duvenaud, David Balduzzi, Devon Hjelm, Fabian Pedregosa, James Martens, Jonathan Lorraine, Julien Perolat and Sören Mindermann for being reviewer for our workshop !

Overview

Advances in generative modeling and adversarial learning gave rise to a recent surge of interest in smooth two-players games, specifically in the context of learning generative adversarial networks (GANs). Solving these games raise intrinsically different challenges than the minimization tasks the machine learning community is used to. The goal of this workshop is to bring together the several communities interested in such smooth games, in order to present what is known on the topic and identify current open questions, such as how to handle the non-convexity appearing in GANs.

Background and objectives

A number of problems and applications in machine learning are formulated as games. A special class of games, smooth games, have come into the spotlight recently with the advent of GANs. In a two-players smooth game, each player attempts to minimize their differentiable cost function which depends also on the action of the other player. The dynamics of such games are distinct from the better understood dynamics of optimization problems. For example, the Jacobian of gradient descent on a smooth two-player game, can be non-symmetric and have complex eigenvalues. Recent work by ML researchers has identified these dynamics as a key challenge for efficiently solving similar problems.

A major hurdle for relevant research in the ML community is the lack of interaction with the mathematical programming and game theory communities where similar problems have been tackled in the past, yielding useful tools. While ML researchers are quite familiar with the convex optimization toolbox from mathematical programming, they are less familiar with the tools for solving games. For example, the extragradient algorithm to solve variational inequalities has been known in the mathematical programming literature for decades, however the ML community has until recently mainly appealed to gradient descent to optimize adversarial objectives.

The aim of this workshop is to provide a platform for both theoretical and applied researchers from the ML, mathematical programming and game theory community to discuss the status of our understanding on the interplay between smooth games, their applications in ML, as well existing tools and methods for dealing with them. We also encourage, and will devote time during the workshop, on work that identifies and discusses open, forward-looking problems of interest to the NeurIPS community.

Responsive image

Call for Contributions



Examples of Topics of Interest to the Workshop:

    We invite researchers to submit work in any related areas, including for example:
  • Other examples of smooth games in machine learning (e.g. actor-critic models in RL).
  • Standard or novel algorithms to solve smooth games. (e.g. Hamiltonian games [Balduzzi et al. 2018])
  • Empirical test of algorithms on GAN applications.
  • Existence and unicity results of equilibria in smooth games.
  • Can approximate mixed equilibria have better properties than the exact ones ? [Arora 2017, Lipton and Young 1994].
  • Variational inequality algorithms [Harker and Pang 1990, Gidel et al. 2018].
  • Handling stochasticity [Hazan et al. 2017] or non-convexity [Grnarova et al. 2018] in smooth games.
  • Related topics from mathematical programming (e.g. bilevel optimization) [Pfau and Vinyals 2016].
  • Please be mindful of the target topics. Topic relevance will be a primordial acceptance criterion.

Submission details

A submission should take the form of an anonymous extended abstract (2-4 pages long excluding references) in PDF format using the following modified NeurIPS style. The submission process will be handled via CMT. Previously published work (or under-review) is acceptable, though it needs to be clearly indicated as published work when submitting. Please provide as a footnote in the actual pdf indicating the venue where the work has been submitted. Submissions can be accepted as contributed talks, spotlight or poster presentations (all accepted submissions can have a poster). Extended abstracts must be submitted by Wednesday October 10, 2018 (11:59pm AoE). Final versions will be posted on the workshop website (and are archival but do not constitute a proceedings).

Tickets

Tickets for the workshop are sold from the NeurIPS website. One ticket per accepted paper has been set aside in a pool of reserved tickets to ensure that the presenters can attend the NeurIPS workshops.

Key Dates:

  • Abstract submission deadline: October 10, 2018 (11:59pm AoE)
  • Acceptance notification: October 29, 2018
  • Camera ready submission: November 15, 2018 (11:59pm AoE)
  • Workshop: December 7, 2018

Invited Speakers

Contributions


Morning Schedule

Time Speaker Title
8:30 Simon Lacoste-Julien
Gauthier Gidel
Opening remarks
Smooth game optimization ?
8:50 Invited talk #1:
Constantinos Daskalakis
-
Improving Generative Adversarial Networks using Game Theory and Statistics [abstract]
9:30 Poster spotlight:
Tianbao Yang
Pavel Dvurechensky
Pavel Dvurechensky
Panayotis Mertikopoulos
-
Hugo Berard
-
Provable Non-Convex Min-Max Optimization.
Solving differential games by methods for finite-dimensional saddle-point problems.
Generalized Mirror Prox Algorithm for Variational Inequalities.
On the convergence of stochastic forward-backward-forward algorithms with variance reduction in pseudo-monotone variational inequalities.
A Variational Inequality Perspective on Generative Adversarial Networks.
10:00 Poster session
10:30 Poster session + Coffee break
11:00 Invited talk #2:
Niao He
-
Smooth Games in Machine Learning Beyond GANs
[abstract]
11:40 Contributed Talk #1:
Volkan Cevher
-
Finding Mixed Nash Equilibria of Generative Adversarial Networks.
12:00 Contributed talk #2:
Pier Giuseppe Sessa
-
Bounding Inefficiency of Equilibria in Continuous Actions Games using Submodularity and Curvature.
12:20 Lunch break

Afternoon Schedule

Time Speaker Title
14:00 Invited talk #3:
Jacob Abernethy
-
Building Algorithms by Playing Games
[abstract]
14:40 Contributed talk #3:
Reyhane Askari Hemmat
-
Negative Momentum for Improved Game Dynamics.
15:00 Coffee break
15:30 Contributed talk #4:
Gabriele Farina
-
Regret Decomposition in Sequential Games with Convex Action Spaces and Losses.
15:50 Invited talk #4:
Paulina Grnarova
-
An interpretation of GANs via online learning and game theory
[abstract]
16:30 Poster spotlight:
Nicolo Fusi
Chidubem Arachie
Joao Monteiro
Steffen Wolf
-
Model Compression with Generative Adversarial Networks.
An Adversarial Labeling Game for Learning from Weak Supervision.
Multi-objective training of Generative Adversarial Networks with multiple discriminators.
A GAN framework for Instance Segmentation using the Mutex Watershed Algorithm.
17:00 Discussion panel:
Simon Lacoste-Julien
Panel with the invited speakers:
Constantinos Daskalakis, Niao He, Jacob Abernethy, Paulina Grnarova
17:30 Organizers Concluding remarks
17:40 Poster session
18:30 Workshop ends

Organizers

Advisors

Relevant References

Abernethy, J.D., Bartlett, P.L., Rakhlin, A., Tewari, A., Optimal strategies and minimax lower bounds for online convex games. In COLT 2009.

Arora, S., Ge, R., Liang, Y., Ma, T., Zhang, Y., Generalization and Equilibrium in Generative Adversarial Nets (GANs). In ICML 2017.

Balduzzi, D., Racaniere, S., Martens, J., Foerster, J., Tuyls, K. and Graepel, T., 2018. The Mechanics of n-Player Differentiable Games. In ICML 2018.

Daskalakis, C., Goldberg, P., Papadimitriou, C., The Complexity of Computing a Nash Equilibrium. SIAM J. Comput., 2009.

Daskalakis, C., Ilyas, A., Syrgkanis, V., Zeng, H., Training GANs with Optimism. In ICLR 2018.

Ewerhart, C., Ordinal Potentials in Smooth Games (SSRN Scholarly Paper No. ID 3054604). Social Science Research Network, Rochester, NY, 2017.

Fedus, W., Rosca, M., Lakshminarayaan, B., Dai, A.M., Mohamed, S., Goodfellow, I., Many Paths to Equilibrium: GANs Do Not Need to Decrease a Divergence At Every Step. In ICLR 2018.

Gidel, G., Jebara, T., Lacoste-Julien, S. Frank-Wolfe Algorithms for Saddle Point Problems. In AISTATS 2017.

Gidel, G., Berard,H., Vincent, P., Lacoste-Julien, S., A Variational Inequality Perspective on Generative Adversarial Networks. arXiv:1802.10551 [cs, math, stat], 2018.

Grnarova, P., Levy, K.Y., Lucchi, A., Hofmann, T., Krause, A., An Online Learning Approach to Generative Adversarial Networks. In ICLR 2018.

Harker, P.T., Pang, J.-S., Finite-dimensinal variational inequality and nonlinear complementarity problems: A survey of theory, algorithms and applications. Mathematical Programming, 1990.

Hazan, E., Singh, K., Zhang, C., Efficient Regret Minimization in Non-Convex Games, in ICML 2017.

Karlin, S., Weiss, G., The Theory of Infinite Games, Mathematical Methods and Theory in Games, Programming, and Economics, 1959.

Lipton, R.J., Young, N.E., Simple Strategies for Large Zero-sum Games with Applications to Complexity Theory, in STOC 94.

Mescheder, L., Nowozin, S., Geiger, A., The Numerics of GANs. In NeurIPS 2017.

Nisan, N., Roughgarden, T., Tardos, E., Vazirani, V., Algorithmic Game Theory. Cambridge University Press, 2007.

Pfau, D., Vinyals, O., Connecting Generative Adversarial Networks and Actor-Critic Methods. arXiv:1610.01945 [cs, stat], 2016.

Roughgarden, T., Intrinsic Robustness of the Price of Anarchy, in: Communications of The ACM - CACM, 2009.

Scutari, G., Palomar, .P., Facchinei, F., Pang, J. s., Convex Optimization, Game Theory, and Variational Inequality Theory. IEEE Signal Processing Magazine, 2010.

Syrgkanis, V., Agarwal, A., Luo, H., Schapire, R.E., Fast Convergence of Regularized Learning in Games, in NeurIPS 2015.

Von Neumann, J., Morgenstern, O., Theory of Games and Economic Behavior. Princeton University Press, 1944.

Address


Place Jean Paul Riopelle, Montreal, QC H2Z 1H5, CANADA

Convention center


Palais des congrès de Montréal