HSE University Presents Research Results at AI Conference in Oman

© HSE University

In April 2026, the International Conference on Intelligent Systems and Artificial Intelligence Applications (ISAA 2026) was held at the University of Nizwa in the Sultanate of Oman. The event was co-organised by HSE University, the University of Nizwa, and the University of Technology and Applied Sciences–Ibri. Researchers from HSE University were among the key speakers at the conference.

The International Conference on Intelligent Systems and Artificial Intelligence Applications (ISAA 2026) provides a unique platform for researchers, academics, and students to engage with industry practitioners and discuss current research on emerging trends in intelligent systems.

The conference featured several tracks, including Artificial Intelligence, Big Data Analytics, Cybersecurity, Intelligent Systems and Robotics, Smart Energy and Smart Cities, Quantum Computing, and Intelligent Systems in Business, Healthcare, and Wireless Networks.

Two scientists from HSE University—Sergey Samsonov, Head of the International Laboratory of Stochastic Algorithms and High-Dimensional Inference at the AI and Digital Science Institute (Faculty of Computer Science), and Sergey Koltsov, Professor in the Department of Informatics at HSE University–St Petersburg, were among the key speakers at the conference.

Sergey Koltsov

Sergey Koltsov made a presentation titled 'From Prompt Engineering to Architectural Clinical Safety: Managing Hallucinations in Medical LLMs.' He discussed the problem of hallucinations in large language models (LLMs) used in medicine—situations in which models generate linguistically plausible but clinically incorrect responses.

According to recent studies, the rate of hallucinations in medical LLMs can reach 15–40%. The professor suggests rethinking how this problem is addressed: rather than relying solely on prompt optimisation or additional model training, he suggests architectural management of reasoning through a multi-level context hierarchy, which includes clinical safety rules, structured patient data, retrieval-augmented generation (RAG) and MemoRAG technologies, as well as dialogue history. Using a clinical decision support system in otorhinolaryngology as an example, he demonstrated a significant reduction in hallucinations and improved sensitivity in screening.

Sergey Samsonov

Sergey Samsonov spoke about modern sampling methods and the current challenges in generating random variables with a given distribution. He reviewed recent trends in the development of MCMC algorithms and amortised sampling methods, particularly diffusion samplers and their relationship to reinforcement learning. He also focused on applications in biological modelling and the structuring of reasoning in language models.

Other representatives of HSE University also took part in ISAA 2026. Hadi Saleh, Associate Professor at the HSE FCS School of Software Engineering, led a tutorial session on automated MLOps platforms for deploying and managing AI services. Using the SmartMLOps platform as an example, participants explored the full lifecycle of AI models—from research prototypes to production solutions, including CI/CD pipelines, automated deployment, monitoring, access control, and performance metrics analysis.

Nikita Morozov, Junior Research Fellow at the Centre of Deep Learning and Bayesian Methods and doctoral student at the HSE Faculty of Computer Science, presented a tutorial on modern generative sampling methods. He also discussed recent developments in MCMC algorithms, generative flow networks (GFlowNets), and diffusion samplers, with an emphasis on their role in generative modelling tasks.

Ekaterina Trofimova and Saraa Ali from the Laboratory of Methods for Big Data Analysis at the HSE AI and Digital Science Institute, lecturers at the HSE Faculty of Computer Science, delivered a two-day introduction to large language models and effective prompting. Participants learned both basic and advanced prompting techniques—from clear, precise formulations to designing complex interaction scenarios—and analysed typical errors while reviewing practical cases from various fields.

According to Evgeniia Kocheva, Director of the HSE Representative Office in the Middle East (Abu Dhabi, UAE), HSE University’s participation in ISAA 2026 marked an important step toward building long-term partnerships with leading universities in the region.

Evgeniia Kocheva

'The conference demonstrated the high level of interest in Russian research in artificial intelligence and machine learning: our colleagues from Oman and other Middle Eastern countries actively engaged with the HSE delegation both during the plenary sessions and at the tutorials. For the HSE Representative Office in Abu Dhabi, this is further confirmation that the university holds a prominent position in the region’s scientific landscape and is ready to pursue joint research projects with partners from the Arab world. For HSE University, co-organising an international conference with a partner university opens up new opportunities to promote the research of our scientists at the global level. We can see serious potential for developing academic exchanges and joint publications with colleagues from the University of Nizwa and other institutions in the region. This is precisely the kind of strategic work for which the HSE Representative Office in the Middle East was established.'