Functions on a convex set which are both $\omega$-semiconvex and $\omega$-semiconcave
V\'aclav Kry\v{s}tof, Lud\v{e}k Zaj\'i\v{c}ek

TL;DR
This paper proves that functions on certain convex sets that are both semiconvex and semiconcave with the same modulus are necessarily $C^{1, ext{omega}}$-smooth, and shows the optimality of the set assumptions.
Contribution
It establishes the optimal conditions under which functions that are both semiconvex and semiconcave are $C^{1, ext{omega}}$-smooth, with simplified proofs and quantitative versions.
Findings
Functions are $C^{1, ext{omega}}$-smooth under optimal convex set conditions.
Provides direct short proofs and quantitative versions of the main result.
Implications for converse Taylor theorems and line-wise smoothness conditions.
Abstract
Let be an open convex set which is either bounded or contains a translation of a convex cone with nonempty interior. It is known that then, for every modulus , every function on which is both semiconvex and semiconcave with modulus is (globally) -smooth. We show that this result is optimal in the sense that the assumption on cannot be relaxed. We also present direct short proofs of the above mentioned result and of some its quantitative versions. Our results have immediate consequences concerning (i) a first-order quantitative converse Taylor theorem and (ii) the problem whether whenever is continuous and smooth in a corresponding sense on all lines. We hope that these consequences are of an independent interest.
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
TopicsOptimization and Variational Analysis · Functional Equations Stability Results · Advanced Optimization Algorithms Research
