The Complexity of Min-Max Optimization for Quadratic Polynomials
Researchers have established that finding approximate stationary points for min-max optimization problems involving quadratic polynomials over a hypercube is PPAD-hard. This complexity holds even for multilinear polynomials with limited variable occurrences and inverse polynomial approximation factors. Consequently, this work presents the first PPAD-hardness results for two-team zero-sum polymatrix games. AI