{"id":"arxiv/2605.17963","url":"https://arxiv.org/abs/2605.17963","tags":["math.OC","stat.ML"],"score":0,"title":"Wasserstein Saddle-Free Newton: Saddle Escape and Local Spectral Convergence","domain":"math.OC","source":"arxiv","authors":["Razvan-Andrei Lascu","Taiji Suzuki"],"reviews":0,"subject":"math.OC","summary":"Extending saddle-free Newton methods to Wasserstein space requires controlling Hessian dependent operators along nonlinear transport trajectories and accounting for a Hessian spectrum that may accumulate at zero. We address these challenges through Wasserstein Saddle-Free Newton (WSFN), a second-order method for nonconvex optimization over Wasserstein space that preconditions the Wasserstein gradient by a regularized inverse square root Hessian operator. The dynamics exploit curvature magnitude while maintaining negative curvature repulsion and positive curvature attraction, avoiding key…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-05-18","reactions":{}}