HMM Speech Recognition with Neural Net Discrimination

Two approaches were explored which integrate neural net classifiers with Hidden Markov Model (HMM) speech recognizers. Both attempt to improve speech pattern discrimination while retaining the temporal processing advantages of HMMs. One approach used neural nets to provide second-stage discrimination following an HMM recognizer. On a small vocabulary task, Radial Basis Function (RBF) and back-propagation neural nets reduced the error rate substantially (from 7.9% to 4.2% for the RBF classifier). In a larger vocabulary task, neural net classifiers did not reduce the error rate. They, however, outperformed Gaussian, Gaussian mixture, and knearest neighbor (KNN) classifiers. In another approach, neural nets functioned as low-level acoustic-phonetic feature extractors. When classifying phonemes based on single 10 msec. frames, discriminant RBF neural net classifiers outperformed Gaussian mixture classifiers. Performance, however, differed little when classifying phones by accumulating scores across all frames in phonetic segments using a single node HMM recognizer. -This work was sponsored by the Department of the Air Force and the Air Force Office of Scientific Research. HMM Speech Recognition with Neural Net Discrimination 195