MMSE-Based Missing-Feature Reconstruction With Temporal Modeling for Robust Speech Recognition | Publicación