Reviews: Causal Inference with Noisy and Missing Covariates via Matrix Factorization

Details: 74: typo'marix' 69-81: bit compact for the large body of related work 97: '.. treatment and control respectively' this suggest 0 treatment, 1 control … 115: perhaps a small figure depicting system model would be nice here … 149: '.. so long as the .. not vanish too fast' given a concrete example of where this is the case 169: MCAR is usually a rather strong/unrealistic assumption in psychometric tests, but ok … 172: '.. exp.family noise mechanism'.