HSE and Yandex Propose Method to Speed Up Neural Networks for Image Generation
A team of scientists at HSE FCS and Yandex Research has proposed a method that reduces computational costs and accelerates text-to-image generation in diffusion models without compromising quality. These models currently set the standard for text-to-image generation, but their use is limited by high computational loads, the company said in a statement.
It is further specified that the developed method—Scale-wise Distillation of Diffusion Models (SwD)—avoids redundant computations during image generation, allowing results to be produced in just 0.3–0.4 seconds.
According to one of the authors, the generation process in diffusion models typically requires dozens of steps involving high-resolution computations. However, in the early stages, only the general structure of the image is formed, and fine details are not yet distinguishable; as a result, some of these computations are redundant. The proposed SwD framework addresses this problem in two ways. First, generation begins at a low resolution and is progressively refined as noise is reduced, eliminating redundant computations in the early stages. Second, the method uses distillation of pre-trained models such as FLUX and Stable Diffusion 3.5, whereby a simpler student model learns to replicate the output of a more complex one, reducing the number of generation steps from dozens to just 4–6.
The authors propose a new loss function for training—Maximum Mean Discrepancy (MMD), which compares how the teacher model 'sees' an image at its internal processing levels with how the student model represents the same image. Unlike traditional approaches, this method does not require auxiliary models, simplifying and accelerating training, the company emphasises. Moreover, MMD can be used as a standalone distillation (knowledge compression) technique: in experiments, the time per training iteration was reduced sevenfold compared to more complex combined approaches.
The new approach reduces generation time from several seconds to just 0.3–0.4 seconds while maintaining visual quality. As a result, SwD makes modern diffusion models faster and more cost-efficient to use, improving their accessibility for practical applications, the company statement says.
The solution is described in a paper to be presented at ICLR 2026, one of the leading conferences on artificial intelligence.
See also:
Personal Interest in Doctoral Thesis Topic Most Important for Confidence in Successful Defence
A researcher at HSE University analysed data on 1,539 doctoral students from 161 Russian universities to identify which features of a thesis topic are associated with academic success and engagement. The most important factor was found to be personal interest in the research topic, which was associated with almost all key aspects of doctoral programme experience—from engaging with the academic supervisor to research activity and confidence about successfully defending the thesis. The findings have been published in Higher Education.
Algebra, Geometry, and AI: Russian and Vietnamese Mathematicians Discuss Current Research
A delegation of scientists from Hanoi visited the HSE Faculty of Computer Science and then took part in a Russian-Vietnamese conference in St Petersburg. The events were part of the three-year project ‘Flexibility and Computational Methods.’ Over the course of the project, the researchers have prepared joint publications and obtained new mathematical results.
Researchers Develop Methodology to Assess the Quality of Legal Representation in Criminal Proceedings
Having a good defence attorney in criminal proceedings can largely determine whether a defendant retains their freedom, health and good name. Researchers at HSE University propose a method for predicting an attorney’s performance based on the outcomes of their previous cases. The methodology takes into account the severity of the charges, the complexity of the cases, and the most likely outcome, drawing on judicial statistics.
When Pictures Hinder Understanding: Illustrations May Impede Learning of Abstract Ideas
Illustrations can help remember specific actions but do not always make abstract ideas easier to learn. Researchers from HSE University and Humboldt University compared how people learn from texts with different levels of abstractness. They found that participants remembered illustrations better and performed better on related tasks after reading a multimedia text about yoga asanas than after reading an abstract text about the Nash equilibrium. The findings could help improve the selection of illustrations for educational and informational materials. The study has been published in Learning and Instruction.
HSE Researchers Present Study on Neural Network Robustness in Video Analysis at ChinaMM 2026
Researchers from HSE University presented a study at the international ChinaMM 2026 conference on how video transmission distortions affect the performance of neural networks. The authors proposed a new approach to testing the robustness of such models under conditions that closely resemble the real-world use of video services.
HSE University to Develop Predictive Analytics System for Icebreaker Motors
Industrial automation is one of the key applications of artificial intelligence. A predictive analytics system for large electric motors is among the solutions being developed for the industry as part of HSE University’s Strategic Technological Project ‘Multi-Agent Platform of AI Solutions for Industry-Specific Tasks.’ What is predictive analytics, how can it improve the operation of electric motors, and what specialists joined forces to develop this technology? Anton Zarubin, Dean of the School of Computer Science, Physics, and Technology at HSE University–St Petersburg and the project development coordinator, explains in this interview with the HSE News Service.
Scientific Expedition to Hainan: HSE Scientists Organise Conference on Statistical AI in China
The Statistical AI Conference was held in Sanya, Hainan Island, China, from 24 to 28 August 2026. The international event brought together leading experts in statistics, machine learning, and applied AI. Alexey Naumov, Director of the AI and Digital Science Institute at the HSE Faculty of Computer Science, and Sergey Samsonov, Head of the International Laboratory of Stochastic Algorithms and High-Dimensional Inference, were among the conference organisers.
How to Assess Students’ Knowledge in the Age of AI
A researcher at HSE University has proposed a flowchart to help lecturers decide how to assess students who use artificial intelligence. It shows where the use of AI should be restricted and where it can be incorporated into the learning process. The article has been published in IT Professional.
5th Fall into ML Conference to Bring Together Leading AI Researchers
The AI and Digital Science Institute at the HSE Faculty of Computer Science invites researchers, developers and everyone shaping the future of technology to the fifth, anniversary edition of the Fall into Machine Learning conference (Fall into ML 2026). The event will take place on October 23–24, 2026, at the HSE Cultural Centre in Moscow and will become the key meeting point for Russia’s AI community.
HSE University to Present Its Projects at the International Youth Festival
The International Youth Festival (IYF) will be held in Yekaterinburg from 11 to 17 September 2026. HSE University will take an active part in the event. The university's booth will present a space settlement mock-up, robots, and the iFORA big data analysis system. Fifty HSE students will visit festival venues, and experts will participate in the business programme.