The Average-Case Time Complexity of Certifying the Restricted Isometry Property

by   Yunzi Ding, et al.

In compressed sensing, the restricted isometry property (RIP) on M × N sensing matrices (where M < N) guarantees efficient reconstruction of sparse vectors. A matrix has the (s,δ)-𝖱𝖨𝖯 property if behaves as a δ-approximate isometry on s-sparse vectors. It is well known that an M× N matrix with i.i.d. 𝒩(0,1/M) entries is (s,δ)-𝖱𝖨𝖯 with high probability as long as s≲δ^2 M/log N. On the other hand, most prior works aiming to deterministically construct (s,δ)-𝖱𝖨𝖯 matrices have failed when s ≫√(M). An alternative way to find an RIP matrix could be to draw a random gaussian matrix and certify that it is indeed RIP. However, there is evidence that this certification task is computationally hard when s ≫√(M), both in the worst case and the average case. In this paper, we investigate the exact average-case time complexity of certifying the RIP property for M× N matrices with i.i.d. 𝒩(0,1/M) entries, in the "possible but hard" regime √(M)≪ s≲ M/log N, assuming that M scales proportional to N. Based on analysis of the low-degree likelihood ratio, we give rigorous evidence that subexponential runtime N^Ω̃(s^2/N) is required, demonstrating a smooth tradeoff between the maximum tolerated sparsity and the required computational power. The lower bound is essentially tight, matching the runtime of an existing algorithm due to Koiran and Zouzias. Our hardness result allows δ to take any constant value in (0,1), which captures the relevant regime for compressed sensing. This improves upon the existing average-case hardness result of Wang, Berthet, and Plan, which is limited to δ = o(1).


page 1

page 2

page 3

page 4


Sparsity and ℓ_p-Restricted Isometry

A matrix A is said to have the ℓ_p-Restricted Isometry Property (ℓ_p-RIP...

Average-case Hardness of RIP Certification

The restricted isometry property (RIP) for design matrices gives guarant...

Johnson-Lindenstrauss Embeddings with Kronecker Structure

We prove the Johnson-Lindenstrauss property for matrices Φ D_ξ where Φ h...

Compressed Sensing and Matrix Completion with Constant Proportion of Corruptions

We improve existing results in the field of compressed sensing and matri...

Distributional Hardness Against Preconditioned Lasso via Erasure-Robust Designs

Sparse linear regression with ill-conditioned Gaussian random designs is...

Non-Convex Compressed Sensing with Training Data

Efficient algorithms for the sparse solution of under-determined linear ...

Uniform Approximations for Randomized Hadamard Transforms with Applications

Randomized Hadamard Transforms (RHTs) have emerged as a computationally ...

Please sign up or login with your details

Forgot password? Click here to reset