Universal Barrier is $n$-Self-Concordant
Yin Tat Lee, Man-Chung Yue

TL;DR
This paper proves that the universal barrier's self-concordance parameter on any n-dimensional convex domain is at most n, providing a tight bound that improves previous results and introduces new moment inequalities for s-concave distributions.
Contribution
It establishes a tight upper bound of n for the self-concordance parameter of the universal barrier on convex domains, improving prior bounds and introducing novel moment inequalities.
Findings
The self-concordance parameter is bounded by n for any n-dimensional convex domain.
The bound is tight, matching the lower limit.
New moment inequalities for s-concave distributions are developed.
Abstract
This paper shows that the self-concordance parameter of the universal barrier on any -dimensional proper convex domain is upper bounded by . This bound is tight and improves the previous bound by Nesterov and Nemirovski. The key to our main result is a pair of new, sharp moment inequalities for -concave distributions, which could be of 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
TopicsSparse and Compressive Sensing Techniques · Mathematical Approximation and Integration · Stochastic Gradient Optimization Techniques
