Sber announced the winners of the international AI competition AI Journey Contest 2023, organized by the bank in collaboration with the AIRI Artificial Intelligence Institute. The prize fund for the contest totaled 11.5 million rubles, underscoring the ambition to push forward practical AI development across industries.
This year’s event drew participation from more than 1,300 data scientists representing 20 countries, including Russia, India, Indonesia, Kazakhstan, and others. A record total of over 12,000 solutions were downloaded as part of the AI Journey Contest 2023, according to the official release. That figure more than doubles the volume seen in the previous year, highlighting a surge of interest and a wide range of approaches being tested on real-world problems.
The contest embraced the SUPER framework as its guiding goalset: Power, Unique, Personalization, Equal, and Recovery. These pillars were reflected in the scale of innovation, the diversity of ideas, and the emphasis on inclusive, broadly beneficial AI outcomes.
Alexander Vedyakhin, First Deputy Chairman of the Board of Directors of Sberbank, highlighted that such competitions serve as a fertile ground for advancing artificial intelligence technologies. He noted that the year produced a record number of decisions, yet he stressed that the quality of solutions mattered more than sheer quantity. Participants presented unconventional and effective methods to tackle challenging issues, demonstrating high professionalism and imaginative problem solving. His remarks emphasized the positive impact these efforts can have on society through responsible AI development.
In addition, a cash prize of one million rubles was awarded for the best scientific paper within the AI Journey 2023 conference collection. A team from Innopolis University and MIPT — Mikhail Rudakov, Alexander Beznosikov, Yaroslav Kholodov, and Alexander Gasnikov — accepted the award for their paper titled “Compression techniques for layer and gradient activations for distributed training of artificial intelligence models.” Their work explored optimization strategies that can accelerate distributed AI training while preserving accuracy, a key concern for scalable AI deployment.
Overall, more than 270 applications were submitted from various countries to nominate the most pressing topics in artificial intelligence research. An expert commission selected 38 articles for inclusion in the scientific collection and planned publication in a special issue of the journal titled Reports of the Russian Academy of Sciences, covering domains such as mathematics, computer science, and management processes.
Reflecting on the program, organizers noted the introduction of a substantial cash prize for the best scientific publication on this topic in the AI Journey track, a gesture aimed at reinforcing the country’s commitment to advancing AI and machine learning research. The prize also supports the dissemination of knowledge through a dedicated scientific collection, reinforcing ties between industry practice and academic inquiry while encouraging further collaboration among researchers, practitioners, and policymakers to accelerate beneficial AI innovations.