MODEL OF SPEECH MONITORING SYSTEM With ARTIFICIAL INTELLIGENCE SYSTEM FOR OF UKRAINIAN LANGUAGE RECOGNITION

Abstract

One of the means of improving speech is control over the use of specific vocabulary - the absence or presence of certain word usage. At one of the stages of functioning of the speech monitoring and analysis system, a necessary step is the user's speech recognition in real-time. During recognition, speech is transcribed, and the result is saved as a text file that can be edited, searched, etc. The Ukrainian language has several unique features that can affect the performance of speech recognition systems, for example, complex phonetics, morphology and syntax. Therefore, the task of developing a speech monitoring system that would transcribe speech in real-time using modern methods and tools is important, and it is advisable to start analyzing relevant methods and developing a model of such a system. A promising direction for solving speech recognition problems is the use of artificial intelligence systems, the use of such systems is becoming increasingly widespread, and their implementation is being improved. Speech recognition is based on acoustic and speech models. Acoustic modeling works out the relationship between linguistic units and audio signals, language modeling is used to distinguish words that sound the same or similar. The proposed speech monitoring and analysis system has among its main functions speech control, where the recognition results are compared with a dictionary, which can be configured to include or exclude specific words, and slang. If a match between the entered text and the dictionary is detected, the app will trigger an alarm that can be sent to the user's device, such as a phone or wristband. The vibration level can be adjusted and can also be ignored for certain words or phrases. External specialists can be involved in filling out such a dictionary, if it is, for example, about control of speech recovery in the event of an injury, or specialists in the education of children with special needs, if the system will be used to control the acquisition of necessary vocabulary. The use of methods and means of artificial intelligence is appropriate for the development of an information system that will be used to analyze speech in Ukrainian to control the quality of vocabulary, in this case, the system should respond to the use of profanity. The use of such an information system (in the form of a smartphone application) will help reduce the use of unwanted words and increase the culture of communication, improving the speech environment.