Current browse context:
math
Change to browse by:
References & Citations
Mathematics > Optimization and Control
Title: Survey Descent: A Multipoint Generalization of Gradient Descent for Nonsmooth Optimization
(Submitted on 30 Nov 2021 (v1), last revised 27 Sep 2022 (this version, v5))
Abstract: For strongly convex objectives that are smooth, the classical theory of gradient descent ensures linear convergence relative to the number of gradient evaluations. An analogous nonsmooth theory is challenging. Even when the objective is smooth at every iterate, the corresponding local models are unstable and the number of cutting planes invoked by traditional remedies is difficult to bound, leading to convergences guarantees that are sublinear relative to the cumulative number of gradient evaluations. We instead propose a multipoint generalization of the gradient descent iteration for local optimization. While designed with general objectives in mind, we are motivated by a ``max-of-smooth'' model that captures the subdifferential dimension at optimality. We prove linear convergence when the objective is itself max-of-smooth, and experiments suggest a more general phenomenon.
Submission history
From: X.Y. Han [view email][v1] Tue, 30 Nov 2021 18:28:17 GMT (1503kb,D)
[v2] Wed, 29 Dec 2021 19:25:52 GMT (1853kb,D)
[v3] Thu, 6 Jan 2022 15:49:23 GMT (1853kb,D)
[v4] Tue, 7 Jun 2022 16:36:04 GMT (3154kb,D)
[v5] Tue, 27 Sep 2022 17:42:31 GMT (3153kb,D)
Link back to: arXiv, form interface, contact.