A Mean-Field Analysis of Neural Stochastic Gradient Descent-Ascent for Functional Minimax Optimization
It provides theoretical guarantees for neural network-based minimax optimization in functional spaces, applicable to domains like policy evaluation and nonparametric regression, but is incremental as it extends mean-field analysis to this setting.
This paper tackles minimax optimization over infinite-dimensional function classes of overparameterized two-layer neural networks, establishing global convergence of the stochastic gradient descent-ascent algorithm to a stationary point at a sublinear rate of O(T^{-1} + α^{-1}) and showing it solves functional equations under strong convexity.
This paper studies minimax optimization problems defined over infinite-dimensional function classes of overparameterized two-layer neural networks. In particular, we consider the minimax optimization problem stemming from estimating linear functional equations defined by conditional expectations, where the objective functions are quadratic in the functional spaces. We address (i) the convergence of the stochastic gradient descent-ascent algorithm and (ii) the representation learning of the neural networks. We establish convergence under the mean-field regime by considering the continuous-time and infinite-width limit of the optimization dynamics. Under this regime, the stochastic gradient descent-ascent corresponds to a Wasserstein gradient flow over the space of probability measures defined over the space of neural network parameters. We prove that the Wasserstein gradient flow converges globally to a stationary point of the minimax objective at a $O(T^{-1} + α^{-1})$ sublinear rate, and additionally finds the solution to the functional equation when the regularizer of the minimax objective is strongly convex. Here $T$ denotes the time and $α$ is a scaling parameter of the neural networks. In terms of representation learning, our results show that the feature representation induced by the neural networks is allowed to deviate from the initial one by the magnitude of $O(α^{-1})$, measured in terms of the Wasserstein distance. Finally, we apply our general results to concrete examples including policy evaluation, nonparametric instrumental variable regression, asset pricing, and adversarial Riesz representer estimation.