Study on recognition of speech based on HMM/MLP hybrid network
Huang Xinye, Ma Xiaohui, Xiang Li, Yuqing Fu, LU Ji-ren · 2002
This paper presents a new HMM/MLP hybrid network for speech recognition. By taking advantage of the discriminative training of MLPs, the unreasonable assumption on the model correctness of ML training in basic HMMs can be overcome, and the discriminative power and recognition performance can be improved. Experimental results demonstrate the efficiency of the HMM/MLP.