On the spectral norm of Rademacher matrices
Rafa{\l} Lata{\l}a

TL;DR
This paper provides non-asymptotic bounds on the spectral norm of Rademacher matrices, confirming a conjecture up to a triple logarithmic factor and improving bounds for matrices with binary coefficients.
Contribution
It establishes bounds on the spectral norm of Rademacher matrices, confirming a conjecture and refining results for matrices with binary entries.
Findings
Conjecture holds up to log log log n factor for general Rademacher matrices.
Triple logarithm factor can be eliminated for matrices with {0,1} coefficients.
Provides non-asymptotic bounds for the spectral norm.
Abstract
We discuss two-sided non-asymptotic bounds for the mean spectral norm of nonhomogenous weighted Rademacher matrices. We show that the recently formulated conjecture holds up to factor for arbitrary Rademacher matrices and the triple logarithm may be eliminated for matrices with -coefficients.
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
TopicsMatrix Theory and Algorithms
