Information theoretic limits of learning a sparse rule

NeurIPS 2020