$\ell_p$-Regression in the Arbitrary Partition Model of Communication

We consider the randomized communication complexity of the distributed $\ellp$-regression problem in the coordinator model, for $p\in (0,2]$. In this problem, there is a coordinator and $s$ servers. The $i$-th server receives $A^i\in{-M, -M+1, \ldots, M}^{n\times d}$ and $b^i\in{-M, -M+1, \ldots, M}^n$ and the coordinator would like to find a $(1+\epsilon)$-approximate solution to $\min{x\in\mathbb{R}^n} |(\sumi A^i)x - (\sumi b^i)|_p$. Here $M \leq \mathrm{poly}(nd)$ for convenience. This model, where the data is additively shared across servers, is commonly referred to as the arbitrary partition model. We obtain significantly improved bounds for this problem. For $p = 2$, i.e., least squares regression, we give the first optimal bound of $\tilde{\Theta}(sd² + sd/\epsilon)$ bits. For $p \in (1,2)$,we obtain an $\tilde{O}(sd^2/\epsilon + sd/\mathrm{poly}(\epsilon))$ upper bound. Notably, for $d$ sufficiently large, our leading order term only depends linearly on $1/\epsilon$ rather than quadratically. We also show communication lower bounds of $\Omega(sd² + sd/\epsilon^2)$ for $p\in (0,1]$ and $\Omega(sd² + sd/\epsilon)$ for $p\in (1,2]$. Our bounds considerably improve previous bounds due to (Woodruff et al. COLT, 2013) and (Vempala et al., SODA, 2020).