The Institute for Computer Science, Artificial Intelligence and Technology (INSAIT) will present a new latest generation artificial intelligence designed to work in Bulgarian. The event will take place on November 19, at 11 am at the Rector’s Office of Sofia University, the press centre of the Ministry of Education and Science said.
The event will be attended by caretaker Prime Minister Dimitar Glavchev, caretaker Minister of Education and Science Prof. Galin Tsokov, and Rector of Sofia University Prof. Georgi Valchev.
During the presentation INSAIT Scientific Director Prof. Martin Vechev and Executive Director Borislav Petrov will present the possible applications of the new artificial intelligence, as well as data on its performance in comparison with some of the most popular products of technology giants from around the world.
INSAIT is the first institute in Eastern Europe to offer world-class working conditions that match those of leading Western universities and institutes. The Institute is structured as a special unit of Sofia University. It was established in partnership with the Swiss ETH Zurich and EPFL, two of the best technological universities in the world. Leading scientists from top American, European and Israeli universities, and research laboratories, guide and advise the institute. INSAIT’s primary focus is on scientific excellence: world-class research, attracting international scientists of outstanding achievement, training the next generation of students and technology leaders.
As BTA reported in January, INSAIT presented BgGPT – the first open large artificial intelligence (AI) language model of the last generation, created for the Bulgarian state, for users, for public and private organizations.
GPT technology is a pinnacle in the development of artificial intelligence, revolutionizing the way society interacts with information and language. As part of INSAIT’s strategy to create open and publicly accessible artificial intelligence for the benefit of society, BgGPT has been specially trained for the specifics of the Bulgarian language, which will enable the creation of applications and solutions to benefit Bulgarian education, business, healthcare, public administration and other fields. In its first version, the model is trained on over 3 billion sentences and consists of 7 billion parameters.