Last week: 1.RNN DAE 2.language vector using PLP features
This week: 1.chinese dataset 2.RNN DAE using data with music noise and reverberation 3.language vector which the input has posterior information
