Fast adaptation of deep neural networks (DNN) is an important research topic in deep learning. Here, we have proposed a general adaptation scheme for DNN based on discriminant condition codes, which are directly fed to various layers of a pre-trained DNN through a new set of connection weights.

References:

[1] S. Xue, O. Abdel-Hamid, H. Jiang, L. Dai, Q. Liu, “Fast Adaptation of Deep Neural Network based on Discriminant Codes for Speech Recognition,” IEEE/ACM Trans. on Audio, Speech and Language Processing, pp.1713-1725, Vol. 22, No. 12, December 2014.

[2] S. Xue, O. Abdel-Hamid, H. Jiang, L. Dai, “Direct Adaptation of Hybrid DNN/HMM Model for Fast Speaker Adaptation,” Proc. of of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP'14), Florence, Italy, May 2014.

[3] O. Abdel-Hamid and H. Jiang, “Rapid and Effective Speaker Adaptation of Convolutional Neural Network Based Models for Speech Recognition,” Proc. of Interspeech, Lyon, France, 2013.

[4] O. Abdel-Hamid and H. Jiang, “Fast Speaker Adaptation Of Hybrid NN/HMM Model for Speech Recognition based on Discriminative Learning of Speaker Code,” Proc. of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP'13), Canada, 2013.