St Petersburg University students and staff contribute to the development of Sber’s AI assistant for universities
At the international conference on artificial intelligence and data analysis, AI Journey 2024, the chatbot project "AI Assistant for Universities", in which mathematicians from St Petersburg University had participated, was presented. The development, initiated by Sber, is designed to support and optimise the educational, research, and organisational activities of higher education institutions.
Students and lecturers of the joint St Petersburg University and Sber academic programme "Artificial Intelligence and Data Science" contributed to the creation of the chatbot. The primary task assigned to the University’s students and staff was to develop modules for the AI assistant based on data from the official website of St Petersburg University.
The AI Journey international conference on artificial intelligence and data analysis serves as a platform for sharing knowledge and experiences in leveraging AI for human benefit. Since 2012, it has brought together AI developers and business representatives from countries including Russia, China, the UAE, India, and Indonesia.
"The chatbot, which operates using a large language model, applies machine learning algorithms and natural language processing. It analyses extensive text data about the university to provide responses to queries from students and staff. However, the capabilities of such assistants are not limited to this; they can potentially handle tasks such as translation, content generation, and process automation," explained Fedor Bushmelev, Assistant Professor in the Department of Informatics at St Petersburg University.
AI Assistant for Universities is a social project by Sber, currently being implemented in Kaliningrad, Krasnodar, Moscow, St Petersburg, and Ulyanovsk. Eight partner universities are participating in the initiative. For instance, a cooperation agreement has been signed with St Petersburg University to jointly implement academic, scientific, and educational projects, as well as contracts for academic programmes.
The developers emphasise that the already available features of the assistant for universities will streamline and automate navigation on the St Petersburg University website. Students, applicants, or employees of the University will be able to find answers to questions about the teaching and learning or administrative processes. For instance, they can ask how to apply for welfare support, secure an increased scholarship, or obtain an academic transcript. The primary advantage of the service is its user-friendly design, allowing users to phrase queries in their natural vocabulary without adhering to strict templates. Typos are not a problem, though offensive words or phrases may result in restricted access.
"Today, most students use messengers, with Telegram being the most popular among them. People are accustomed to communicating in chats, so we anticipate that the AI Assistant for Universities will become a convenient service where users can pose questions similar to those they might ask an academic office staff member, admissions committee representative, or even a groupmate," said Fedor Bushmelev.
St Petersburg University is in the top 5 Russia’s universities for training IT specialists, as recognised by the Russoft 2024 Award. As part of a grant project supported by the Russian Ministry of Science and Higher Education, six academic programmes on artificial intelligence have been developed at St Petersburg University in collaboration with partners. Among them are the following: Mathematical Robotics and Artificial Intelligence; Artificial Intelligence and Big Data Science; Big Data and Distributed Digital Platform; Artificial Intelligence and Data Science (bachelor’s and master’s programmes); and Artificial Intelligence in Speech Modelling. Staff from St Petersburg University have also contributed to the development of the first textbook on artificial intelligence for schoolchildren.
In the near future, the development team aims to optimise the new Timetable module and refine it to accommodate a wider range of user queries. As part of their research, students from the Artificial Intelligence and Data Science academic programme will also compile a list of the most frequently asked questions to enhance response times and improve the extraction of information from complex tables and documents.