This article describes how to use heterogeneous information in speech enhancement. In most of the current speech enhancement systems, clean speeches are recovered only from the signals collected by acoustic microphones, which will be greatly affected by the acoustic noises. However,...