‘The Competition Gave Young Researchers an Opportunity to Take the Initiative’

© iStock

In September, HSE University announced the results of a competition of digital projects by early-career HSE scientists. The event was organised within the framework of the strategic project ‘Digital Transformation: Technologies, Effects, Efficiency’. The organisers selected 8 out of 22 applications. The research teams have already started to implement their projects, and the results will be presented at the end of November. The HSE News Service shares the details of three of the highest-scoring projects in the competition. The creators of the projects are staff members of the HSE Center for Language and Brain, MIEM, and the Faculty of Computer Science.

‘The Rey Test: Development of a Multilingual Digital Application for the Diagnosis of Memory Disorders in Children’

Olga Buivolova, junior research fellow at the HSE Center for Language and Brain, heads this project. The main focus of her research is speech disorders in adults: her PhD thesis, which she defended last year, is devoted to linguistic tests for the diagnosis of such disorders. In addition, her research interests include memory disorders. Together with her colleagues from the HSE Center for Language and Brain, she participated in the development of a Russian-language version of the test by Swiss psychologist André Rey to detect memory disorders in adults.

Now, a group of researchers of the Center for Language and Brain (including research assistants, students, and doctoral students) led by Olga Buivolova has started to develop a similar test to detect memory disorders in children. An important feature of this project is that the test will be offered not only in Russian, but also in four other languages common in Russia. Versions in Adyghe and Ukrainian have already been finalised, while Tatar and Yakut versions are in development. The researchers receive translation support from volunteers who are native speakers of these languages.

To work with the test, the researchers developed a digital application that gives the user step-by-step instructions. During testing, the child answers various questions The answers help researchers to determine how well the child remembers information, how quickly they forget it, what is the maximum capacity of their memory, etc. To determine the typical range of performance of each task, the researchers collect empirical data; a survey of school children has already started, and about 100 people will be interviewed in two months.

Olga Buivolova, junior research fellow at the HSE Center for Language and Brain, project head

 ‘The collected data will make the developed tool standardised and reliable. It is user-friendly and can be used not only by specialists in clinical practice, but also by teachers and even parents. Scientists will use the application to observe how memory develops in children.’

‘Digital Technology for the Acquisition and Subsequent Verification of Experimental Data on the Electrification of Polymer Dielectrics of Space Technology’

This project was submitted to the competition by Ilshat Mullahmetov (project manager) and Semyon Tolstikov, early-career scientists of the HSE Doctoral School of Engineering Science studying at the MIEM School of Electronic Engineering. The researchers are graduates of the Master’s Programme ‘Computer Systems and Networks’ with a specialisation in IT. During their studies at the MIEM Laboratory of Space Vehicles and Systems' Functional Safety, they became interested in space-related topics. Their master's theses, which they successfully defended, were also dedicated to space, and now they are working on their dissertations for a candidate of sciences degree.

They are continuing their research in the field of space materials science as part of their project. When designing a particular technological device for operation in space, it is necessary to determine what will happen to the materials it’s made of: what effects will occur, whether some of the effects will be unfavourable, and if they occur, what needs to be done to prevent them. Answering these questions involves intensive laboratory work, but it is still not possible to fully recreate what happens in space.

The digital technology developed by Ilshat Mullahmetov and Semyon Tolstikov will help to obtain the necessary data on polymer materials (dielectrics) and verify conclusions without conducting additional time-consuming experiments. The technology can be applied during space technology design and testing. The project will result in a computer programme that will help space explorers in the future.

Semyon Tolstikov, Doctoral Student, School of Electronic Engineering

'The HSE Digital Transformation Project Competition for early-career scientists is interesting because ‘it gives young researchers an opportunity to take the initiative at the junction of their main research work and what is in demand now in the field of digital technologies for our university and for the country in general.’

‘A Reparametrization Trick for Differentiable Rendering’

Kirill Struminsky, research fellow at the Centre of Deep Learning and Bayesian Methods at the HSE Faculty of Computer Sciences, heads this project. He is a mathematician specialising in machine learning in graduate school. As a researcher, he is most interested in problems at the intersection of mathematics and computer science. Since childhood, he has been fond of computer graphics. ‘It’s amazing how accurately you can transfer the reality around you to a computer screen,’ says Struminsky.

Differentiable rendering is a new field at the intersection of computer graphics, computer vision and machine learning that requires a knowledge of mathematics. As such, Kirill Struminsky's project fully corresponds to his research interests. He works with a group of colleagues, including doctoral students and master’s students, whose theses he supervised. 

The project is dedicated to multi-view reconstruction—when a three-dimensional model of a particular scene is built based on a set of flat projections (for example, photographs). This task has traditionally been solved using geometric means and classical computer vision methods, but due to objective limitations, these could not achieve high fidelity or realism. A couple of years ago, it was found that for a more successful solution, it is necessary to describe not only the object geometry, but also the features of light propagation. Machine learning can solve this problem.

Kirill Struminsky, Research Fellow, Centre of Deep Learning and Bayesian Methods

‘Machine learning helps us to not only accurately determine the geometry, but also predict how light works from new angles—that is, to fully describe the dynamics of light in the entire scene. The project proposes a modern algorithm for calculating a physical model, which focuses on the parts that can be calculated faster or slower.’

The researcher believes that this approach will help to improve the quality of three-dimensional representations of the surrounding world. The new technology can be used in the reconstruction of objects, interior design, in medicine, etc.