RESEARCH AND DEVELOPMENT OF A METHOD FOR SYTHESIZING SENTENCES IN THE KAZAKH LANGUAGE BASED ON MACHINE LEARNING

Abstract

Today, sentence synthesis is used in various fields. These are voice assistants, IVR systems, smart homes, chatbots and much more. Some time ago, machine learning appeared in the field of speech synthesis, as in many other fields. Machine learning is a broad set of artificial intelligence that studies methods for creating algorithms capable of learning. It turned out that a number of components of the entire system can be replaced by neural networks, which allows not only to approach the existing algorithms with quality, but even significantly surpass them. The article provides an overview of sentence synthesis technologies, solves the problem of sentence synthesis in the Kazakh language based on a chatbot system using the seq2seq method. A parallel corpus of questions and answers has been collected in the Kazakh language. The corpus of questions and answers in Kazakh was collected as a result of translation and cleaning of many corpora, such as Cornell movie, Ubantu and others, which are used to create many chatbots in English. A number of experiments were carried out and results were obtained using corpora based on the constructed model for the synthesis of sentences in the Kazakh language.