Development of a virtual voice assistant to fulfill user requests using AI
Students Name: Denys Andrii Romanovych
Qualification Level: magister
Speciality: Information Technology Design
Institute: Institute of Computer Science and Information Technologies
Mode of Study: full
Academic Year: 2024-2025 н.р.
Language of Defence: англійська
Abstract: Denys A.R., Denysiuk P.Yu. (thesis supervisor). Development of a virtual voice assistant to fulfill user requests using AI. Master qualification work. “Lviv Polytechnic National University”, Lviv, 2024. Extended annotation In this explanatory note, graphic and textual information about the entire necessary process of familiarization with the subject and the implementation of its practical part is quite extensively described. In general, the structure of the note reached 100+ pages, where three sections are described, introduction, conclusions, etc., in which there are graphic elements. The largest number of illustrations is used in the third chapter, where the practical part is described, 28 pictures are illustrated. In the first chapter, there is a table comparing the available voice assistants. The second chapter also has illustrations and one graph. More than 25 sources were used, which in turn helped to better understand the chosen topic, both with a practical component and in general for personal understanding. The note also contains an appendix that stores some of the code responsible for the query execution functionality of this software application, as well as 7 sheets of graphics containing various flowcharts, UML diagrams, and more. The use of a voice assistant has always been and will be relevant in the use of the surrounding society. This topic has been studied by professionals more than once, and even for dozens of years, in order to be able to carry out certain improvements of all the necessary components. Of course, in the future, this functionality will make life much easier for ordinary users, as this technology is rapidly developing. Since the dynamics of life are gradually increasing, ordinary gadget users do not have time to read long posts on the Internet, write text, watch long videos, etc. That is why a voice assistant comes to the rescue here, according to the forecasts of leading media resources, in the near future, about half of the requests on the Internet will be used through a voice assistant, which already indicates a remarkable result of the use of artificial intelligence [2]. Voice assistants make it possible to minimize, and in some cases completely eliminate, the need to use peripheral devices, and even hands or eyes to view content on the Internet. In this work, a detailed introduction and analysis of this subject was carried out, however, one should not forget about the negative side. The main problem is privacy, since not all programs have good protection. If we take as an example the world leader in the use of voice assistants - Alexa, then she could send messages to the wrong addressees, and it is good that it is only a message, and not some important information. That is why it is worth making some improved corrections, or ideas on how to better optimize work, or on the contrary, minimize it to begin with, using proven basic commands [1]. In the course of this thesis, the creation of a program is described, the essence of which is to execute commands from received voice requests from the user. First of all, the product must satisfy the potential user with a set of commands, be clear during use, and sometimes have a simple graphical interface, in turn, further updating of the software application is planned in order to develop the product to attract potential users, etc. During the implementation of the practical part of the thesis, the Python programming language was used to create the product, with all the necessary and additionally downloaded libraries, the main component in this work is a separately downloaded library - Speech Recognition. Thanks to this library, in fact, almost all part of the work is done from the software code, since, using certain modules, the user connects his sound devices and is able to communicate with artificial intelligence, etc. More details about the practical part can be found in the third section of this master’s thesis. All existing and future queries are stored in a MySQL database, which can also be connected to Python code through certain third-party libraries. As for the graphical representation for the end user, several options were worked out in theory, regarding the future implementation, etc. However, one of the most optimal options turned out to be the creation of a simple WEB page, through which requests will be sent to our server part, thus, the site will not be loaded with enough graphics, everything will be presented briefly and clearly, the start button, introductory text, instructions. Accordingly, it was decided to use the Django framework to connect the frontend for our application, and in turn minimal skills of web tools are involved - HTML, CSS, JS. During execution, the following development environments were also used: PyCharm and Visual Studio Code. Keywords: voice assistant, artificial intelligence, Python, Django, Siri, graphical interface, end user, voice, request, frontend, backend. List of used sources: Методи та системи штучного інтелекту: теорія та практика // Булгакова О.С., Зосімов В.В., Поздєєв В.О. // Інформаційні технології – 2020 // [Навчальний посібник] Research on the Development of Voice Assistants in the Era of Artificial Intelligence [Електронний ресурс] / Yuqi Huang // Syntaxxx. – 2022. – Режим доступу до ресурсу: https://www.shs-conferences.org/articles/shsconf/pdf/2023/04/shsconf_sdmc2022_03019.pd Humanizing voice assistant: The impact of voice assistant personality on consumers’ attitudes and behaviors // Atieh Poushneh // Journal of Retailing and Consumer Services – 2021. [Електронний ресурс] – Режим доступу до ресурсу: https://www.sciencedirect.com/science/article/pii/S0969698920312911