AI Institute Professor Represents RTU MIREA at NeurIPS 2025 in Mexico City

In 2024–2025, the RTU MIREA team of two Institute of Artificial Intelligence students — Arseny Sychev and Arseny Lenin — joined the international nonprofit CayleyPy project.​

Their research as part of the project proposed a new ML-based method for finding the shortest paths on ultralarge-scale Cayley graphs. It showed record-setting results against benchmarks for Rubik’s cubes and other similar complex puzzles with up to 10^145 nodes.​

The paper describing the research was accepted to one of the most prestigious AI conferences, NeurIPS 2025 (A* category) and was listed as Spotlight, a tag awarded to less than 15% of the accepted contributions. Its preprint is accessible on arXiv.​

Professor Aleksei Mikhailovich Romanov of the AI Institute is presenting the research, along with taking part in a poster session, at NeurIPS in Mexico City this week. He has already given talks at Mexico’s largest university, UNAM, and at CINVESTAV (the Center for Research and Advanced Studies of the National Polytechnic Institute).​

Keep it up!​