gaussian-matrices

This paper identifies sufficient sample sizes for maximum-likelihood support recovery of sparse binary signals in high-SNR regimes, revealing an information-theoretic threshold and trade-offs between measurement sparsity and computational cost.