cs.LGMay 19, 2026

Optimal Reconstruction from Linear Queries

Authors: Yuval FilmusShay MoranElizaveta Nesterova

Organizations: Technion – Israel Institute of Technology · Google Research

Abstract

We study the problem of reconstructing an unknown point in Rd\mathbb{R}^d from approximate linear queries. This setting arises naturally in applications ranging from low-dimensional remote sensing and signal recovery to high-dimensional data analysis and privacy-sensitive inference. Our main goal is to characterize the optimal reconstruction error as a function of the number of queries TT, the ambient dimension dd, and the noise parameter δδ. We first analyze the limit TT \to \infty and show that the optimal reconstruction error converges to the explicit value 2d/(d+1)δ\sqrt{2d/(d+1)} δ, which plays a role analogous to the Bayes optimal error in supervised learning. When the dimension is fixed, we show that the excess error above this limit decays doubly exponentially fast as TT \to \infty, a rate that is significantly faster than those typically encountered in learning curves. When the dimension grows, we show that a number of queries on the order of exp(d)\exp(d) is necessary and sufficient to achieve vanishing excess error. Finally, we introduce and analyze an improper variant of the reconstruction problem. From a technical perspective, our main contribution is a generalization of Jung's theorem (1901). The classical theorem bounds the maximum possible radius of a set of diameter 1 and characterizes extremal bodies. Our generalization provides a robust variant that characterizes near-extremal bodies and is proved via geometric and dynamical arguments exploiting symmetry and Lie group actions.

Explore similar work

May 13, 2026cs.DS

Min-Max Optimization Requires Exponentially Many Queries

We study the query complexity of min-max optimization of a nonconvex-nonconcave function ff over [0,1]d×[0,1]d[0,1]^d \times [0,1]^d. We show that, given oracle access to ff and to its gradient f\nabla f, any algorithm that finds an ε\varepsilon-approximate stationary point must make a number of queries that is exponential in 1/ε1/\varepsilon or dd.
Martino Bernasconi, Matteo Castiglioni, Andrea Celli +1
Feb 18, 2026math.ST

Separating Oblivious and Adaptive Models of Variable Selection

Sparse recovery is among the most well-studied problems in learning theory and high-dimensional statistics. In this work, we investigate the statistical and computational landscapes of sparse recovery with \ell_\infty error guarantees. This variant of the problem is motivated by \emph{variable selection} tasks, where the goal is to estimate the support of a kk-sparse signal in Rd\mathbb{R}^d. Our main contribution is a provable separation between the \emph{oblivious} (for each'') and \emph{adaptive} (for all'') models of \ell_\infty sparse recovery. We show that under an oblivious model, the optimal \ell_\infty error is attainable in near-linear time with klogd\approx k\log d samples, whereas in an adaptive model, k2\gtrsim k^2 samples are necessary for any algorithm to achieve this bound. This establishes a surprising contrast with the standard 2\ell_2 setting, where klogd\approx k \log d samples suffice even for adaptive sparse recovery. We conclude with a preliminary examination of a \emph{partially-adaptive} model, where we show nontrivial variable selection guarantees are possible with klogd\approx k\log d measurements.
Ziyun Chen, Jerry Li, Kevin Tian +1
May 2, 2026cs.LG

Linear-Readout Floors and Threshold Recovery in Computation in Superposition

Two recent approaches to computation in superposition reach different recursive capacity regimes: Hänni et al. certify O~(d3/2)\tilde{O}(d^{3/2}) computable features in width dd via an approximate-linear recursive template, while Adler and Shavit reach near-quadratic capacity (up to logarithmic factors) using thresholded Boolean recovery. The main contribution of this paper is conceptual: we argue these results are not contradictory because they maintain different interface invariants, and we formalize the distinction. As a tool, we record a rank-trace Welch-type lower bound for biorthogonal linear readouts: for FdF \gg d, the worst-case off-diagonal cross-talk of any unit-diagonal linear readout is Ω(d1/2)Ω(d^{-1/2}), and the bound is tight on average for unit-norm tight frames. At quadratic feature load F=d2F=d^2, random-support threshold recovery succeeds for sparsities s=O(d/logd)s=O(d/\log d), while linear readouts still incur Ω(s/d)Ω(s/d) average per-coordinate squared error on Bernoulli sparse states. Matching the Welch floor against the published tolerance of the Hänni correction layer explains the d3/2d^{3/2} scale as a compatibility threshold for that template, not a universal upper bound. Robust nonlinear reset beyond the Hänni template is left open.
Hector Borobia, Elies Seguí-Mas, Guillermina Tormo-Carbó