https://papers.cool/venue/criscitiello22a@v178@PMLR
Negative curvature obstructs acceleration for strongly geodesically convex optimization, even with...
Hamilton and Moitra (2021) showed that, in certain regimes, it is not possible to accelerate Riemannian gradient descent in the hyperbolic plane if we restrict...
convex optimizationnegativecurvatureacceleration