SPEECH RECOGNITION SYSTEM IN KAZAKH LANGUAGE

Abstract

Machine learning is one of the most popular and developing fields in the modern world, and it has an increasing impact on people’s daily life. It is one of the many study methods designed to teach computers without the external help. One of the most important inventions in the field of machine learning is the Automatic Speech Recognition (ASR). ASR is implemented in almost every smartphone, and they are used as voice assistants. The main objective of this project is to design an effective speech recognition system that will be able to understand Kazakh language and convert it to text based on a machine learning algorithm. It was decided to choose this topic because despite the fact that the field of machine learning is developing rapidly, there have not been any ASRs designed yet in the country that could be used on a daily basis. Nowadays, smart devices have become an integral part of majority people’s lives, and the effects of voice assistants are also increasing due to the high development of machine learning. However, these technologies are only available in widely spoken languages, such as Chinese, English, Russian, and Spanish. Therefore, the main aim of this project is to design ASR for specifically Kazakh population.