19
12

Near-Optimality of Linear Recovery from Indirect Observations

Abstract

We consider the problem of recovering linear image BxBx of a signal xx known to belong to a given convex compact set X{\cal X} from indirect observation ω=Ax+ξ\omega=Ax+\xi of xx corrupted by random noise ξ\xi with finite covariance matrix. It is shown that under some assumptions on X{\cal X} (satisfied, e.g., when X{\cal X} is the intersection of KK concentric ellipsoids/elliptic cylinders, or the unit ball of the spectral norm in the space of matrices) and on the norm \|\cdot\| used to measure the recovery error (satisfied, e.g., by p\|\cdot\|_p-norms, 1p21\leq p\leq 2, on Rm{\mathbf{R}}^m and by the nuclear norm on the space of matrices), one can build, in a computationally efficient manner, a "presumably good" linear in observations estimate, and that in the case of zero mean Gaussian observation noise, this estimate is near-optimal among all (linear and nonlinear) estimates in terms of its worst-case, over xXx\in {\cal X}, expected \|\cdot\|-loss. These results form an essential extension of those in our paper arXiv:1602.01355, where the assumptions on X{\cal X} were more restrictive, and the norm \|\cdot\| was assumed to be the Euclidean one. In addition, we develop near-optimal estimates for the case of "uncertain-but-bounded" noise, where all we know about ξ\xi is that it is bounded in a given norm by a given σ\sigma. Same as in arXiv:1602.01355, our results impose no restrictions on AA and BB. This arXiv paper slightly strengthens the journal publication Juditsky, A., Nemirovski, A. "Near-Optimality of Linear Recovery from Indirect Observations," Mathematical Statistics and Learning 1:2 (2018), 171-225.

View on arXiv
Comments on this paper