Embeddings of Schatten Norms with Applications to Data Streams

Abstract

Given an $n \times d$ matrix $A$, its Schatten-$p$ norm, $p \geq 1$, is defined as $\|A\|_p = \left (\sum_{i=1}^{\textrm{rank}(A)}\sigma_i(A)^p \right )^{1/p}$, where $\sigma_i(A)$ is the $i$-th largest singular value of $A$. These norms have been studied in functional analysis in the context of non-commutative $\ell_p$-spaces, and recently in data stream and linear sketching models of computation. Basic questions on the relations between these norms, such as their embeddability, are still open. Specifically, given a set of matrices $A^1, \ldots, A^{\operatorname{poly}(nd)} \in \mathbb{R}^{n \times d}$, suppose we want to construct a linear map $L$ such that $L(A^i) \in \mathbb{R}^{n' \times d'}$ for each $i$, where $n' \leq n$ and $d' \leq d$, and further, $\|A^i\|_p \leq \|L(A^i)\|_q \leq D_{p,q} \|A^i\|_p$ for a given approximation factor $D_{p,q}$ and real number $q \geq 1$. Then how large do $n'$ and $d'$ need to be as a function of $D_{p,q}$? We nearly resolve this question for every $p, q \geq 1$, for the case where $L(A^i)$ can be expressed as $R \cdot A^i \cdot S$, where $R$ and $S$ are arbitrary matrices that are allowed to depend on $A^1, \ldots, A^t$, that is, $L(A^i)$ can be implemented by left and right matrix multiplication. Namely, for every $p, q \geq 1$, we provide nearly matching upper and lower bounds on the size of $n'$ and $d'$ as a function of $D_{p,q}$. Importantly, our upper bounds are {\it oblivious}, meaning that $R$ and $S$ do not depend on the $A^i$, while our lower bounds hold even if $R$ and $S$ depend on the $A^i$. As an application of our upper bounds, we answer a recent open question of Blasiok et al. about space-approximation trade-offs for the Schatten $1$-norm, showing in a data stream it is possible to estimate the Schatten-$1$ norm up to a factor of $D \geq 1$ using $\tilde{O}(\min(n,d)^2/D^4)$ space.