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.

An overview of maximal distance minimizers problem

Inverse maximal and average distance minimizer problems

Asymptotics of maximum distance minimizers

Minimal and Maximal Distances in Metric Spaces

Distance Majorization and Its Applications

The Complexity of Constrained Min-Max Optimization

Distance metric learning by minimal distance maximization

On Well-posed Mutually Nearest and Mutually Furthest Point Problems in Banach Spaces

SION'S MINIMAX THEOREM IN GEODESIC METRIC SPACES AND A RIEMANNIAN EXTRAGRADIENT ALGORITHM

Boundedness for maximal operators over hypersurfaces in $\mathbb{R}^3$

Max-Distance Sparsification for Diversification and Clustering

Minkowski's successive minima in convex and discrete geometry

Variational problems concerning length distances in metric spaces

Ahlfors regularity of continua that minimize maxitive set functions

A minimax approach to duality for linear distributional sensitivity testing

Max-Min Distance Analysis by Using Sequential SDP Relaxation for Dimension Reduction

A non-convex relaxed version of minimax theorems

Existence and regularity of min-max anisotropic minimal hypersurfaces

Statistical Mechanics of Min-Max Problems

Tight Bounds for the Maximum Distance Over a Polytope to a Given Point

On Maximizing the Distance to a Given Point over an Intersection of Balls II