Adaptive regularization with cubics on manifolds

Naman Agarwal, Nicolas Boumal, Brian Bullins, Coralia Cartis

Abstract: Adaptive regularization with cubics (ARC) is an algorithm for unconstrained, non-convex optimization. Akin to the trust-region method, its iterations can be thought of as approximate, safe-guarded Newton steps. For cost functions with Lipschitz continuous Hessian, ARC has optimal iteration complexity, in the sense that it produces an iterate with gradient smaller than $$\varepsilon $$ ε in $$O(1/\varepsilon ^{1.5})$$ O ( 1 / ε 1.5 ) iterations. For the same price, it can also guarantee a Hessian with smallest eigenvalue larger than $$-\sqrt{\varepsilon }$$ - ε . In this paper, we study a generalization of ARC to optimization on Riemannian manifolds. In particular, we generalize the iteration complexity results to this richer framework. Our central contribution lies in the identification of appropriate manifold-specific assumptions that allow us to secure these complexity guarantees both when using the exponential map and when using a general retraction. A substantial part of the paper is devoted to studying these assumptions—relevant beyond ARC—and providing user-friendly sufficient conditions for them. Numerical experiments are encouraging.