Learning Heteroscedastic Models by Convex Programming under Group Sparsity | Read Paper on Bytez