Riemannian ascent--descent for nonconvex nonconcave minimax landscapes: convergence to basin saddle points and applications to distributionally robust optimization

• URL: http://arxiv.org/abs/2609.14141v1
• Date: Sat, 12 Sep 2026 20:42:11 GMT
• Title: Riemannian ascent--descent for nonconvex nonconcave minimax landscapes: convergence to basin saddle points and applications to distributionally robust optimization
• Abstract summary: We study minimax problems over a gradient of a geodesicean space and a Buresian manifold.<n>Because no globally convergent order first method is known to be available, we instead introduce the notion of a emphbasin saddle point.
• Score: 1.0195618602298684
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We study a class of distributionally robust optimization (DRO) problems for the statistical risk problem, formulated as minimax problems over the product of a Euclidean space and a Riemannian manifold. Because the resulting minimax landscape is nonconvex nonconcave in general, no globally convergent first order method is known to be available. We instead introduce the notion of a \emph{basin saddle point}, a Nash equilibrium defined locally on the Cartesian product of a $δ$ basin around a connected component of the local minima critical set and a geodesic ball on the measure manifold. We develop an abstract convergence framework for a Riemannian gradient ascent multistep descent iteration to a basin saddle point under a local Łojasiewicz type growth condition, with exponent $β\in (1,2]$, in the $δ$ basin around connected components of the local minima critical sets. Under Lipschitz regularity of critical sets we establish linear convergence for $β= 2$ and polynomial convergence for $β\in (1,2)$ to a basin saddle point, with explicit dependence on the sectional curvature of the manifold. We then instantiate this framework for the statistical risk DRO problem over Gaussian measures, where the ambiguity set is naturally modeled as the product of Euclidean space and the Bures Wasserstein manifold of covariance matrices, which we relax to a penalized DRO formulation. We derive nonasymptotic Hessian estimates for the resulting Lagrangian, establish existence and local uniqueness of its maximizer, and prove that an alternating Riemannian gradient scheme converges to a basin saddle point of the penalized DRO problem, recovering the linear and polynomial rates of the abstract theory with all constants explicit in terms of data dimension, loss moments, and the reference covariance.

Related papers

• Curvature-Independent Regret Bounds for Distributed Online Optimization on Hadamard Manifolds [9.035866986287205]
  This work addresses decentralized online convexian optimization on Hadamard.<n>To our knowledge these are the first curvature-independent regret guarantees on embeddings for predicted rates, with no degradation due to curvature.
  arXiv  Detail & Related papers  (2026-09-12T01:53:49Z)
• Nonconvex Composite Functional Constraints via First-Order Augmented Lagrangian Methods under Local Regularity [10.643433352418915]
  We show uniformly augmented condition bounds of primal non-balancing non-concave minimax reformulation.<n>With dual regularization, a global local argument together with a bias.<n>$KKT.
  arXiv  Detail & Related papers  (2026-07-09T21:28:29Z)
• Fitting Unknown Number of Hyperplanes with Manifold Optimization [57.48093263119306]
  Fitting unknown number of linearplanes to data is a fundamental yet challenging problem in machine learning.<n>Existing approaches often struggle with optimum optima or lack geometric consistency.
  arXiv  Detail & Related papers  (2026-05-27T14:02:20Z)
• From Saddle Points Toward Global Minima: A Newton-Type Method on Wasserstein Space [51.56484100374058]
  We propose a second-order method that preconditions Wasserstein minimization by a regularized square root of the squared Wasserstein Hessian.<n>This construction preserves attraction toward directions toward directions toward positive curvature while inducing repulsion directions along negative curvature.<n>We show that WSFN converges linearly in $L2$-Wasserstein to a non-de global distance.
  arXiv  Detail & Related papers  (2026-05-18T07:18:22Z)
• Local LMO: Constrained Gradient Optimization via a Local Linear Minimization Oracle [51.714334316332476]
  Local L is a new projection-free type for constrained optimization.<n>Local LMO is perhaps best viewed as an oracle of Gradient Descent (GD)
  arXiv  Detail & Related papers  (2026-05-09T10:03:24Z)
• Riemannian Flow Matching for Disentangled Graph Domain Adaptation [51.98961391065951]
  Graph Domain Adaptation (GDA) typically uses adversarial learning to align graph embeddings in Euclidean space.<n>DisRFM is a geometry-aware GDA framework that unifies embedding and flow-based transport.
  arXiv  Detail & Related papers  (2026-01-31T11:05:35Z)
• On lower bounds of the density of planar periodic sets without unit distances [55.2480439325792]
  We introduce a novel approach to estimating $m_1(mathbbR2)$ by reformulating the problem as a Maximal Independent Set (MIS) problem on graphs constructed from flat torus.<n>Our experimental results, supported by theoretical justifications of proposed method, demonstrate that for a sufficiently wide range of parameters this approach does not improve the known lower bound.
  arXiv  Detail & Related papers  (2024-11-20T12:07:19Z)
• Riemannian stochastic optimization methods avoid strict saddle points [68.80251170757647]
  We show that policies under study avoid strict saddle points / submanifolds with probability 1. This result provides an important sanity check as it shows that, almost always, the limit state of an algorithm can only be a local minimizer.
  arXiv  Detail & Related papers  (2023-11-04T11:12:24Z)
• Extragradient Type Methods for Riemannian Variational Inequality Problems [25.574847201669144]
  We show that the average-rate convergence of both REG and RPEG is $Oleft(frac1Tright)$, aligning with in the 2020ean case. Results are enabled by addressing judiciously the holonomy effect so additional observations can be reduced.
  arXiv  Detail & Related papers  (2023-09-25T14:08:02Z)
• Last-Iterate Convergence of Adaptive Riemannian Gradient Descent for Equilibrium Computation [52.73824786627612]
  This paper establishes new convergence results for textitgeodesic strongly monotone games.<n>Our key result shows that RGD attains last-iterate linear convergence in a textitgeometry-agnostic fashion.<n>Overall, this paper presents the first geometry-agnostic last-iterate convergence analysis for games beyond the Euclidean settings.
  arXiv  Detail & Related papers  (2023-06-29T01:20:44Z)
• Decentralized Weakly Convex Optimization Over the Stiefel Manifold [28.427697270742947]
  We focus on the Stiefel manifold in the decentralized setting, where a connected network of agents in $nMn log-1)$ are tested. We propose an method called the decentralized subgradient method (DRSM)$ for forcing a natural station below $nMn log-1)$.
  arXiv  Detail & Related papers  (2023-03-31T02:56:23Z)
• Randomized Coordinate Subgradient Method for Nonsmooth Composite Optimization [11.017632675093628]
  Coordinate-type subgradient methods for addressing nonsmooth problems are relatively underexplored due to the set of properties of the Lipschitz-type assumption.
  arXiv  Detail & Related papers  (2022-06-30T02:17:11Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.