Abstract:Despite its important applications in Machine Learning, min-max optimization of nonconvex-nonconcave objectives remains elusive. Not only are there no known first-order methods converging even to approximate local min-max points, but the computational complexity of identifying them is also poorly understood. In this paper, we provide a characterization of the computational complexity of the problem, as well as of the limitations of first-order methods in constrained min-max optimization problems with nonconvex-nonconcave objectives and linear constraints. As a warm-up, we show that, even when the objective is a Lipschitz and smooth differentiable function, deciding whether a min-max point exists, in fact even deciding whether an approximate min-max point exists, is NP-hard. More importantly, we show that an approximate local min-max point of large enough approximation is guaranteed to exist, but finding one such point is PPAD-complete. The same is true of computing an approximate fixed point of Gradient Descent/Ascent. An important byproduct of our proof is to establish an unconditional hardness result in the Nemirovsky-Yudin model. We show that, given oracle access to some function $f : P \to [-1, 1]$ and its gradient $\nabla f$, where $P \subseteq [0, 1]^d$ is a known convex polytope, every algorithm that finds a $\varepsilon$-approximate local min-max point needs to make a number of queries that is exponential in at least one of $1/\varepsilon$, $L$, $G$, or $d$, where $L$ and $G$ are respectively the smoothness and Lipschitzness of $f$ and $d$ is the dimension. This comes in sharp contrast to minimization problems, where finding approximate local minima in the same setting can be done with Projected Gradient Descent using $O(L/\varepsilon)$ many queries. Our result is the first to show an exponential separation between these two fundamental optimization problems.

Optimality Conditions and Complexity for Non-Lipschitz Constrained Optimization Problems

First-order Optimality Conditions for Two Classes of Generalized Nonsmooth Semi-Infinite Optimization

New second-order optimality conditions for directional optimality of a general set-constrained optimization problem

Optimality conditions for invex nonsmooth optimization problems with fuzzy objective functions

Characterizing Parametric and Convergence Stability in Nonconvex and Nonsmooth Optimizations: A Geometric Approach

The Complexity of Constrained Min-Max Optimization

Necessary conditions for the optimal control of a shape optimization problem with non-smooth PDE constraints

Optimality conditions for the nonlinear programming problems on Riemannian manifolds

Optimality Conditions of Strict Minimality in Optimization Problems under Inclusion Constraints.

Convexification, Concavification and Monotonization in Global Optimization

Efficiency Conditions for Nonsmooth Vector Equilibrium Problems with Constraints via Generalized Subdifferentials

Second-Order Strong Optimality and Second-Order Duality for Nonsmooth Constrained Multiobjective Fractional Programming Problems

Convex and Non-convex Optimization Under Generalized Smoothness

Norm-induced Cuts: Optimization with Lipschitzian Black-box Functions

Directional Compactness, Approximations and Efficiency Conditions for Nonsmooth Vector Equilibrium Problems with Constraints

A Note on Complexity for Two Classes of Structured Non-Smooth Non-Convex Compositional Optimization

Optimal control of a non-smooth elliptic PDE with non-linear term acting on the control

Second-order optimality conditions for general nonconvex optimization problems and variational analysis of disjunctive systems

Generalized Derivatives and Optimality Conditions in Nonconvex Optimization

On the directional asymptotic approach in optimization theory

Generalized Gradient Projection Neural Networks for Nonsmooth Optimization Problems