Fast $(1+\varepsilon)$-Approximation Algorithms for Binary Matrix Factorization