Vocabulary independent speech recognition using particles

A method is presented for performing speech recognition that is not dependent on a fixed word vocabulary. Particles are used as the recognition units in a speech recognition system which permits word-vocabulary independent speech decoding. A particle represents a concatenated phone sequence. Each string of particles that represents a word in the one-best hypothesis from the particle speech recognizer is expanded into a list of phonetically similar word candidates using a phone confusion matrix. The resulting word graph is then re-decoded using a word language model to produce the final word hypothesis. Preliminary results on the DARPA HUB4 97 and 98 evaluation sets using word bigram redecoding of the particle hypothesis show a WER of between 2.2% and 2.9% higher than using a word bigram speech recognizer of comparable complexity. The method has potential applications in spoken document retrieval for recovering out-of-vocabulary words and also in client-server based speech recognition.