Study of deep learning and CMU sphinx in automatic speech recognition

Machine learning has proven to be a very effective tool in automatic speech recognition. This paper is an attempt to give a broad overview of the applications of various approaches of machine learning in speech recognition with special reference to deep learning and CMU Sphinx. Deep learning in Speech recognition is a relatively recent development. On the other hand, CMU Sphinx, an open source software has been in use for this purpose for a relatively longer time. CNN, a Deep Learning algorithm learns the invariant features that help it to differentiate between different words and word sequences. CMU Sphinx uses GMM-HMM model to predict the phonemes in the utterance to determine the word or set of continuous words that were spoken.