The winners of the poster and project competitions at the SMILES 2024 Summer School on Machine Learning, which concluded at the historic Kochubey Mansion in Tsarskoe Selo near St. Petersburg, have been announced. Students who successfully completed the intensive 12-day course also received certificates signed by Skoltech President Alexander Kuleshov and Professor Evgeny Burnaev, who heads the AI and Applied AI centers at Skoltech. 

“This is the fifth time we’ve organized SMILES,” pointed out SMILES 2024 Director and Skoltech Assistant Professor Alexey Zaytsev. “Our goal is not only to teach students the latest methods in deep learning, but also to develop their overall scientific skills. This includes mastering methods, reading and writing scientific papers, and effectively communicating with colleagues. We aim to not just educate but to demonstrate how it all works in practice. For this reason, we’ve invited the most engaging and insightful lecturers to the school. I hope the students enjoyed participating and will be able to apply the knowledge and contacts they gained with us in the future. We’ve done everything we can to ensure the school was as useful as possible for them, in a friendly and creative atmosphere. I’m confident that they, along with their colleagues, classmates, and friends, will want to join us again next year.”  

Poster Competition Winners:  

— Ivan Butakov with his work titled “Mutual Information Estimation via Diffeomorphic Transformations and Lossy Compression.” 

— Artem Kharinaev with his research titled “Neural Network Analysis of News in Compliance.” 

— Matvey Anoshin with the poster titled “Hybrid Quantum Cyclic Generative Adversarial Network for Small Molecule Generation.”  

Project competition (hackathon) winners — “AI Methods for Solving Applied and Research Problems” (onsite participants):  

— First place: “Benchmarking Quantization Methods for LLMs” by Ksenia Studenikina, Mikhail Bykov, Omar Zoloev, Artyom Kharinaev. 

— Second place: “Applying Topological Data Analysis Methods to Detect Hallucinations in Language Models” by Irina Zabaryanskaya, Andrey Sukhov, Amina Miftakhova, Stefan Maria Ayluro.

— Third place: “RAG with Additional Information on Embedding Quality” by Mikhail Menshchikov, David Kagramanyan, Anna Zhukovskaya, Alexander Kharitonov.  

Special Nomination from the school’s general sponsor, Sber: “Adversarial Attacks on Transformer Models for Time Series” by Dmitry Anikin, Alexander Nikolaev, Grigory Gerasimov, Matvey Anoshin.  

Hackathon winners (online participants):  

— First place: “Improving Voice Conversion via Extreme Neural Optimal Transport” by Sergey Kudryashov, Vladislav Gromadsky, Oleg Rasin, Mikhail Varekha. 

— Second place: “Pre-training Neural Networks for Medical 3D Image Processing” by Svetlana Danilova, Ivan Betev, Elizaveta Alekseevskaya. 

— Third place: “Detecting Fraud in Chess” by Ustin Zolotikov, Snezhana Erashova, Egor Korshikov.  

Special Nomination from the SMILES 2024 gold sponsor, Alfa Bank: “Neural Network Approach to Credit Scoring” by Artem Sytkin, Maxim Igitov, Nikita Gulyakov, and Mikhail Malyutin.