A study of machine learning algorithms in speech recognition and language identification system

Abstract

Speech recognition is a broad topic that primarily involves sub-topics like language identification, speaker identification, speech emotion recognition, speech to text systems, text to speech systems, dialogue systems and much more. While, human beings are quickly able to recognize or identify a language because of the corpora of knowledge built over the years. However, it is a challenging task to have a machine identify a spoken language. So, to build a system that can correctly identify multiple languages irrespective of the dialect and speaker characteristics is an interesting area of research. One benefit of such a LID system is that the barrier between people caused due to language differences will be broken. Such a system will further the progress of globalization. The latest developments of machine learning in speech and language are described as a detailed state of the art in this paper.