DSA Faculty
API
← к списку преподавателей

Деркач Денис Александрович

Факультет компьютерных наук

Профиль на hse.ru ↗ тел.: 27266
Публикаций
226
Языков
3
Наград
6
Конференций
0
Профиль Публикации (226) Курсы (9)

Профессиональные интересы

физика высоких энергийанализ данных

Должности

  • Заведующий лабораторией — Факультет компьютерных наук, Институт искусственного интеллекта и цифровых наук, Научно-учебная лаборатория методов анализа больших данных
  • Доцент — Факультет компьютерных наук, Департамент больших данных и информационного поиска
  • Научный руководитель образовательной программы — Умные устройства: аппаратная разработка

Био

  • · Начал работать в НИУ ВШЭ в 2015 году.
  • · Научно-педагогический стаж: 7 лет.

Образование

  • 2010 · PhD: Университет Париж XI
  • 2007 · Магистратура: Санкт-Петербургский государственный университет, специальность «Физика», квалификация «Магистр»
  • 2004 · Бакалавриат: Санкт-Петербургский государственный университет, специальность «Физика», квалификация «Бакалавр»

Опыт работы

  • · 2017: н. в

Награды и поощрения

  • · Медаль "Признание - 10 лет успешной работы" НИУ ВШЭ (сентябрь 2025)
  • · Благодарность первого проректора НИУ ВШЭ (август 2024)
  • · Благодарность НИУ ВШЭ (май 2024)
  • · Почетная грамота Министерства науки и высшего образования Российской Федерации (ноябрь 2022)
  • · Благодарность первого проректора НИУ ВШЭ (август 2022)
  • · Благодарность Факультета компьютерных наук НИУ ВШЭ (август 2018)

Гранты и проекты

  • — · на соискание учёной степени кандидата наук

Идентификаторы исследователя

Публикации (226)

Application of Physics-Informed Neural Networks for Solving the Inverse Advection-Diffusion Problem to Localize Pollution Sources

2025 · CHAPTER · en

Abstract. This paper investigates the application of Physics-Informed Neural Networks (PINNs) for solving the inverse advection-diffusion problem to localize pollution sources. The study focuses on optimizing neural network architectures to accurately model pollutant dispersion dynamics under diverse conditions, including scenarios with weak and strong winds and multiple pollution sources. Various PINN configurations are evaluated, showing the strong dependence of solution accuracy on hyperparameter selection. Recommendations for efficient PINN configurations are provided based on these comparisons. The approach is tested across multiple scenarios and validated using real-world data that accounts for atmospheric variability. The results demonstrate that the proposed methodology achieves high accuracy in source localization, showcasing the stability and potential of PINNs for addressing environmental monitoring and pollution management challenges under complex weather conditions.

PDF ↗

Unsupervised Learning for Calorimeter Response Correction: A WGAN-Based Method

2025 · CHAPTER · en

The long-term stability of calorimeters is crucial in high-energy physics experiments, where precise energy measurements are essential for accurate particle reconstruction. This study introduces a Wasserstein GAN (WGAN)-based machine learning approach for calibrating calorimeter responses affected by aging and other systematic shifts. Our methodology is applied to realistic, high-granularity calorimeter data that more accurately mimic physical detector conditions. The dataset reflects energy deposition across all calorimeter cells, following an exponential energy spectrum and eliminating artificial peaks in the distribution. By leveraging Wasserstein distance minimization, our model estimates aging coefficients of cells, realigning degraded detector responses with their undamaged counterparts. The results highlight the potential of a data-driven approach for calorimeter calibration, demonstrating correcting energy measurement discrepancies with a reduced number of required events, making it a valuable tool for future detector calibration strategies.

DOI ↗ PDF ↗

Hybrid Fault Detection in Three-Phase Induction Motors

2025 · CHAPTER · en

Three-phase induction motors play a crucial role in industrial applications due to their efficiency, durability, and reliability. However, effective fault detection remains challenging, primarily due to the scarcity of labeled failure data, which limits the performance of traditional machine learning (ML)-based diagnostic models and increases the risk of overfitting and poor generalization. Conventional methods, such as current signature analysis (CSA), have long been used for motor diagnostics, but can be further enhanced by integrating advanced ML techniques. To address these challenges, we propose a hybrid approach that combines CSA with a ResNet-based deep learning model, incorporating a physically informed synthetic anomaly generation process. This method leverages the predictive capabilities of supervised ML while maintaining the diagnostic robustness of unsupervised signature analysis, resulting in higher accuracy and improved generalization in different motor conditions. Experimental evaluations demonstrate that our approach outperforms traditional ML diagnostic techniques, making it an effective solution for industrial applications. The findings underscore the potential impact of this method in development of intelligent fault detection systems, paving the way for more reliable and automated predictive maintenance strategies in industrial settings.

DOI ↗

Approaches to Proxy Modeling of Gas Reservoirs

2025 · ARTICLE · en

In the gas industry, accurate forecasting of gas production is critical for optimizing well operating conditions. Although traditional hydrodynamic models offer high accuracy, they are often computationally intensive and time-consuming, prompting a growing interest in proxy-based alternatives. This study proposes a hybrid methodology based on Spatio-Temporal Graph Neural Networks (ST-GNNs) for gas production forecasting. The methodology integrates graph neural networks to account for spatial interdependencies between wells with recurrent and convolutional neural networks for time-series analysis. The model was validated using an extensive set of hydrodynamic simulation calculations and real-world field data. On average, the ST-GNN method reduces computational time by a factor of 4.3 compared to traditional hydrodynamic models, with a median predictive error not exceeding 10% across diverse datasets, despite variability in specific scenarios. The ST-GNN framework demonstrates promising potential as a tool for operational and strategic planning.

DOI ↗ PDF ↗

Роль научно-просветительской деятельности в развитии газовых технологий с использованием искусственного интеллекта

2025 · ARTICLE · ru

В статье представлен совместный молодежный научно-­просветительский проект ПАО «Газпром» и вузов-­партнеров «Зайти в АйТи». Обоснована актуальность решения производственных задач кросс-­функциональными командами и необходимость формирования смежных и междисциплинарных компетенций на стыке инженерных специальностей и цифровых технологий.

Observation of the Λ0b → J/ψΞ−K + and Ξ0b → J/ψΞ−π+ decays

2025 · ARTICLE · en

The first observation of the Ξ0b→J/ψΞ−π+ decay and the most precise measurement of the branching fraction of the Λ0b→J/ψΞ−K+ decay are reported, using proton-proton collision data from the LHCb experiment collected in 2016--2018 at a centre-of-mass energy of 13~TeV, corresponding to an integrated luminosity of 5.4~fb−1. Using the Λ0b→J/ψΛ and Ξ0b→J/ψΞ− decays as normalisation channels, the ratios of branching fractions are measured to be: B(Λ0b→J/ψΞ−K+)B(Λ0b→J/ψΛ)=(1.17±0.14±0.08)×10−2, B(Ξ0b→J/ψΞ−π+)B(Ξ0b→J/ψΞ−)=(11.9±1.4±0.6)×10−2, where the first uncertainty is statistical and the second systematic.

DOI ↗ PDF ↗

Search for 𝐷0 meson decays to 𝜋+𝜋−𝑒+𝑒− and 𝐾+𝐾−𝑒+𝑒− final states

2025 · ARTICLE · en

A search for 𝐷0 meson decays to the 𝜋+⁢𝜋−⁢𝑒+⁢𝑒− and 𝐾+⁢𝐾−⁢𝑒+⁢𝑒− final states is reported using a sample of proton-proton collisions collected by the LHCb experiment at a center-of-mass energy of 13 TeV, corresponding to an integrated luminosity of 6 fb−1. The decay 𝐷0 →𝜋+⁢𝜋−⁢𝑒+⁢𝑒− is observed for the first time when requiring that the two electrons are consistent with coming from the decay of a 𝜙 or 𝜌0/𝜔 meson. The corresponding branching fractions are measured relative to the 𝐷0 →𝐾−⁢𝜋−⁢[𝑒+⁢𝑒−]𝜌0/𝜔 decay, where the two electrons are consistent with coming from the decay of a 𝜌0 or 𝜔 meson. No evidence is found for the 𝐷0 →𝐾+⁢𝐾−⁢𝑒+⁢𝑒− decay and world-best limits are set on its branching fraction. The results are compared to, and found to be consistent with, the branching fractions of the 𝐷0 →𝜋+⁢𝜋−⁢𝜇+⁢𝜇− and 𝐷0 →𝐾+⁢𝐾−⁢𝜇+⁢𝜇− decays recently measured by LHCb and confirm lepton universality at the current precision.

DOI ↗ PDF ↗

Test of Lepton Flavor Universality with 𝐵0𝑠 →𝜙ℓ+ℓ− Decays

2025 · ARTICLE · en

Lepton flavor universality in rare 𝑏 →𝑠 transitions is tested for the first time using 𝐵0𝑠 meson decays. The measurements are performed using 𝑝⁢𝑝 collision data collected by the LHCb experiment between 2011 and 2018, corresponding to a total integrated luminosity of 9 fb−1. Branching fraction ratios between the 𝐵0𝑠→𝜙⁢𝑒+⁢𝑒− and 𝐵0𝑠 →𝜙⁢𝜇+⁢𝜇− decays are measured in three regions of dilepton mass squared, 𝑞2, with 0.1

DOI ↗ PDF ↗

Search for B∗0(s)→μ+μ− in B+c→π+μ+μ− decays

2025 · ARTICLE · en

A search for the very rare B∗0→μ+μ− and B∗0s→μ+μ− decays is conducted by analysing the B+c→π+μ+μ− process. The analysis uses proton-proton collision data collected with the LHCb detector between 2011 and 2018, corresponding to an integrated luminosity of 9\,fb−1. The signal signatures correspond to simultaneous peaks in the μ+μ− and π+μ+μ− invariant masses. No evidence for an excess of events over background is observed for either signal decay mode. Upper limits at the 90% confidence level are set on the branching fractions relative to that for B+c→J/ψπ+ decays, B∗0(μ+μ−)π+/J/ψπ+

DOI ↗ PDF ↗

First Determination of the Spin-Parity of Ξ𝑐(3055)+,0 Baryons

2025 · ARTICLE · en

The Ξ0⁢(−)𝑏→Ξ𝑐⁢(3055)+(0)⁢(→𝐷+(0)⁢Λ)⁢𝜋− decay chains are observed, and the spin-parity of Ξ𝑐⁢(3055)+(0) baryons is determined for the first time. The measurement is performed using proton-proton collision data at a center-of-mass energy of √𝑠 =13 TeV, corresponding to an integrated luminosity of 5.4 fb−1, recorded by the LHCb experiment between 2016 and 2018. The spin-parity of the Ξ𝑐⁢(3055)+(0) baryons is determined to be 3/2+ with a significance of more than 6.5⁢𝜎 (3.5⁢𝜎) compared to all other tested hypotheses. The up-down asymmetries of the Ξ0⁢(−)𝑏→Ξ𝑐⁢(3055)+(0)⁢𝜋− transitions are measured to be −0.92±0.10±0.05 (−0.92±0.16±0.22), consistent with maximal parity violation, where the first uncertainty is statistical and the second is systematic. These results support the hypothesis that the Ξ𝑐⁢(3055)+(0) baryons correspond to the first 𝐷-wave 𝜆-mode excitation of the Ξ𝑐 flavor triplet.

DOI ↗ PDF ↗

Курсы (9)