Translation. Region: Russian Federal
Source: Peter the Great St. Petersburg Polytechnic University –
An important disclaimer is at the bottom of this article.
At the end of September, a delegation from SPbPU visited Blagoveshchensk, Russia's Far East. Polytechnic University representatives—professors Alexey Filimonov, Vladimir Zaborovsky, and Vadim Korablyov, and associate professor Vyacheslav Bondarenko of the Higher School of Engineering Physics—participated in several scientific events.
The visit began with a working meeting at the Vostochny Cosmodrome. The delegation toured the technical and refueling complexes, the Soyuz-2 launch pad, and the newly constructed launch pad for the Angara heavy-lift rocket.
The following day, the 23rd scientific conference "Physics: Fundamental and Applied Research, Education" opened. This year, it was held at Amur State University, which celebrated its 50th anniversary. Participants presented key results of experimental and theoretical research in the fields of semiconductor physics, condensed matter, and nanotechnology to the scientific community. Over 50 papers were presented at the conference, including the usual 50 poster presentations.
The Polytechnic University team presented three papers. The first, "Dynamics and Kinetics of Lead Magnoniobat Relaxor," was devoted to the study of ferroelectric relaxor materials with potential for practical application.
Relaxors represent one of the most interesting groups of disordered compounds. In this study, we analyzed structural relaxation in the lead magnoniobate relaxor PbMg1/3Nb2/3O3 (PMN). X-ray photon correlation spectroscopy (XPCS) was employed as the primary method for studying slow dynamic processes. Using experimental data, we traced the temperature evolution of both single-time and dual-time correlation functions. Ultra-broadband dielectric spectroscopy was also used to track the kinetics of the dielectric response of lead magnoniobate during aging in the region of a diffuse phase transition. It was found that aging is accompanied by a hardening and narrowing of the dielectric loss spectra and a decrease in the dielectric strength. An explanation is proposed based on the concept of creating degenerate polar nanoregions spanning several chemically ordered regions.
The second report was titled "Physical Aspects of Machine Learning Processes." It discussed, from the perspective of modern computer science and theoretical physics, the evolution of digital implementations of deep artificial neural networks toward the creation of multimodal transformers of large language models—the foundation of intelligent technologies for modeling complex physical processes and "learning" computer systems.
Pythagoras developed the theory of the harmonic series, which explains why music, like geometry, is a form of reflection of the objective properties of physical reality. A theory capable of guiding the development of artificial intelligence (AI) systems based on fundamental physical concepts has not yet been created. This paper presents an exo-intelligence extension of the architecture of modern computer-based software automata, which it proposes to consider as information-open physical systems capable not only of inductive learning based on explicitly specified digital data but also of conceptual learning. This allows us to solve the problem of regularizing the generative hallucinations of large linguistic models by exploiting fundamental physical laws.
The third report, "Chaotic Potential on the Surface of Doped III-Nitrides," focused on obtaining information about the nature of the electronic properties of semiconductor surfaces and contact structures. SPbPU scientists presented the results of a study of the size effect in semiconductor heterojunctions during space charge distribution across point and extended linear defects, which is relevant for fine-tuning the manufacturing technology of modern electronic devices based on heterojunctions.
This paper discusses the screening of electroactive defects (point and linear) and the structure of the chaotic potential on the GaN surface under self-compensation conditions. Using a statistical analysis of a Poisson ensemble of charged defects, the amplitude and scale of the chaotic potential are determined. It is shown that at high degrees of self-compensation, inhomogeneities in the fields of charged dislocations dominate the surface.
The conference featured an informal scientific discussion with colleagues from Moscow State University, TUSUR, Kabardino-Balkarian State University, Novosibirsk State University, and many other educational and scientific centers in Russia, as well as with representatives from Heihe University in China.
Professor Zaborovsky also delivered a lecture entitled "Information Intervention in Physics: Computer Science and the Problem of Machine Learning" to third- and fourth-year students at the Institute of Computer Science and Engineering at Amur State University as part of an exchange program.
Please note: This information is raw content obtained directly from the source. It represents an accurate account of the source's assertions and does not necessarily reflect the position of MIL-OSI or its clients.