### Current browse context:

math.NA

### Change to browse by:

### References & Citations

# Mathematics > Numerical Analysis

# Title: Comparison of Matrix Norm Sparsification

(Submitted on 30 Jan 2022 (v1), last revised 7 Aug 2022 (this version, v2))

Abstract: A well-known approach in the design of efficient algorithms, called matrix sparsification, approximates a matrix $A$ with a sparse matrix $A'$. Achlioptas and McSherry [2007] initiated a long line of work on spectral-norm sparsification, which aims to guarantee that $\|A'-A\|\leq \epsilon \|A\|$ for error parameter $\epsilon>0$. Various forms of matrix approximation motivate considering this problem with a guarantee according to the Schatten $p$-norm for general $p$, which includes the spectral norm as the special case $p=\infty$.

We investigate the relation between fixed but different $p\neq q$, that is, whether sparsification in the Schatten $p$-norm implies (existentially and/or algorithmically) sparsification in the Schatten $q\text{-norm}$ with similar sparsity. An affirmative answer could be tremendously useful, as it will identify which value of $p$ to focus on. Our main finding is a surprising contrast between this question and the analogous case of $\ell_p$-norm sparsification for vectors: For vectors, the answer is affirmative for $p<q$ and negative for $p>q$, but for matrices we answer negatively for almost all $p\neq q$. In addition, our explicit constructions may be of independent interest.

## Submission history

From: Shay Sapir [view email]**[v1]**Sun, 30 Jan 2022 16:44:59 GMT (18kb)

**[v2]**Sun, 7 Aug 2022 12:49:34 GMT (20kb)

Link back to: arXiv, form interface, contact.