Curves of Descent
D. Drusvyatskiy, A.D. Ioffe, A.S. Lewis

TL;DR
This paper provides a new existence proof for curves of near-maximal slope in nonsmooth optimization and characterizes these curves for semi-algebraic functions as solutions to subgradient dynamical systems.
Contribution
It introduces a transparent existence proof for curves of near-maximal slope and links them to subgradient dynamical systems for semi-algebraic functions.
Findings
Existence of curves of near-maximal slope established.
Characterization of these curves as solutions to subgradient systems for semi-algebraic functions.
Enhances understanding of steepest descent in nonsmooth optimization.
Abstract
Steepest descent is central in variational mathematics. We present a new transparent existence proof for curves of near-maximal slope --- an influential notion of steepest descent in a nonsmooth setting. We moreover show that for semi-algebraic functions --- prototypical nonpathological functions in nonsmooth optimization --- such curves are precisely the solutions of subgradient dynamical systems.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Differential Equations and Dynamical Systems · Advanced Topology and Set Theory · History and Theory of Mathematics
