NSU Students Took Part in the International Machine Learning School SMILES 2026 in China

Students of Novosibirsk State University participated in the international summer school on machine learning SMILES 2026, which took place from July 13 to 27 on the campus of Nanjing University in Suzhou (China). The school was organized by Nanjing University and the Skolkovo Institute of Science and Technology.

A total of 104 students, postgraduate students, and young researchers from Russia, China, and other countries took part in the in-person program. Interest in the school for 2026 reached a record level: the organizers received 2544 applications, and 585 candidates advanced to the extended selection stage.

The selection process for the school was rigorous — according to the organizers, there were more than ten applications per spot. The procedure included several stages: writing a motivation letter, attaching a CV, passing a qualifying test, completing a project of choice, and creating a video in English in which one had to explain a complex topic from the field of artificial intelligence in the simplest language possible.

Nikita Abramov, a 3rd-year student in the field of Mechatronics and Robotics Institute of Intelligent Robotics, NSU, chose the classification of hallucinations in neural networks for the qualifying project, and in the video explained the principle of operation of generative adversarial neural networks (GAN). Based on all stages, participants received a status: offline participation, reserve, or online format.

The core of the two-week program consisted of lectures by leading professors in the field of ML/AI and work on a research project. The school’s lecture program is one of the most international in its history: 17 leading researchers from seven countries, including Russia, China, the United Kingdom, the USA, Saudi Arabia, the UAE, and Vietnam.

The central theme of the school is engineering artificial intelligence and the transition from fundamental methods to reliable solutions for science and industry. Participants studied Bayesian machine learning, probabilistic modeling and uncertainty estimation, large language and generative models, medical AI, computer vision, and intelligent engineering systems.

The project work was focused on scientific research: participants were given hypotheses to test and then defend the results. The work was carried out exclusively in teams.

— I had a project — LLM Implementation in Residue Number System, which means replacing matrix multiplications in neural network layers with a residue class system and subsequently optimizing this process. It was necessary either to prove that this is effective and faster or to prove that there is no gain in memory or speed, although accuracy is maintained in any case. The most difficult part, I would say, was that I was the most experienced team member and had to take on the role of team lead — properly distributing tasks and helping the team understand material that was new to them. Overcoming the language barrier was also challenging: the school was in English, but at the start there were big problems with communication. And the translator was sometimes completely inaccurate, — Nikita Abramov said.

In addition to the educational part, the organizers provided a rich cultural program: team-building games were held for the participants, a football match was organized between the teams of Russia and China, and excursions in Suzhou were arranged — to the Suzhou History Museum and Tiger Hill. One day was allocated for independent exploration of Shanghai: participants were transported to the city by organized transfer in the morning and picked up in the evening, while during the day everyone explored the metropolis on their own.

For Nikita Abramov, participating in SMILES 2026 was his first experience in an international-level school, although he already had experience with international trips (he took part in the World Youth Festival in 2025).

— To understand the topic in such a short time, on top of overcoming a language barrier, explaining complex subjects to foreigners, and trying to understand them — these are definitely the teamwork skills that are valued everywhere at the highest level., — he noted. — And useful contacts were made, that is undeniable.

The NSU delegation also included Polina Karitskaya, a graduate student of the Faculty of Information Technology at NSU 2026, and Anastasia Krohina, a 2023 graduate of the Faculty of Natural Sciences at NSU.

SMILES (Summer School on Machine Learning) is held by Skoltech for the seventh time. In 2025, the school was held outside of Russia for the first time — at the Harbin Institute of Technology, and in 2026, the partner was Nanjing University, a member of the C9 League — an alliance of nine leading research universities in China.