arXiv subjects

Denis Derkach

Publications and source records attributed to Denis Derkach.

At least 19 recordsLinked to original sources

arXiv2026

Evolutionary Two-Stage Hyperparameter Optimization Strategies for Physics-Informed Neural Networks

Physics-Informed Neural Networks (PINNs) solve Partial Differential Equations (PDEs) by embedding physical laws into neural network training. However, their performance suffers from unstable convergence, training plateaus, and strong sensitivity to architectural and optimization hyperparameters due to the highly non-convex and multi-term structure of the physics-informed loss. In this setting, the outer-loop hyperparameter search is a noisy and black-box optimization problem over heterogeneous parameters, where classical local or gradient-based strategies are easily trapped in suboptimal regions. Evolutionary algorithms, with their population-based exploration and ability to handle mixed, non-differentiable search spaces, provide a more robust mechanism for discovering promising configurations. We propose and investigate a two-stage approach based on evolutionary algorithms that combines exploration and exploitation parts of PINNs training to improve solution accuracy and robustness under fixed computational budgets. In the first stage, we perform low-fidelity training runs with truncated epochs to rapidly screen candidate configurations, treating hyperparameter selection as a black-box outer-loop problem. In the second stage, only the most promising candidates are fully trained with standard gradient-based optimizers to refine the solution. Evaluated on three popular problems, namely Advection, Klein-Gordon and Helmholtz equations, our method consistently outperforms standard training and achieves significantly lower mean error within constrained computational resources.

cs.LG↗

arXiv2026

Correcting Sensor-Induced Distribution Drift with Wasserstein Adversarial Learning

The quality of recorded data depends on the stability of the sensor system that acquires it. Sensor motion and aging can degrade the performance and stability of downstream data-driven methods. We present a Wasserstein-GAN-inspired approach for unsupervised inference of physically interpretable transformation parameters that map a changed detector response distribution back to a nominal reference distribution. In contrast to standard generative modeling, the generator is used as a learnable calibration transformation whose trainable weights represent the sought parameters, while the critic provides a distributional distance signal via the Wasserstein objective. We validate the approach on a tracking-detector toy model with controlled layer shifts and demonstrate its application on high-granularity Geant4-simulated calorimeter data with cell-wise aging effects. The method recovers aging coefficients for individual cells with correlation to ground truth and improves agreement between calibrated and reference energy-sum distributions, while exhibiting the expected degradation at increasing channel-to-channel noise levels. These results indicate that adversarial distribution matching can serve as a data-driven component of calibration strategies in settings where direct labels for degradation parameters are unavailable.

cs.LG↗

arXiv2026

Physics-Constrained Neural Networks for Improved Short-Term Weather Forecasting: A Case Study over the South Pacific

This study introduces enhancements to physics-constrained neural networks (PCNNs) that improve the accuracy and stability of hybrid short-term weather forecasting models. Building on the WeatherGFT architecture, three innovations are proposed. First, an upgraded numerical solver, combining a fifth-order weighted essentially non-oscillatory scheme (WENO-5), a beta-plane approximation, and subgrid-scale viscosity, permits a fourfold increase in the integration time step to 1200 s while reducing the daily mean squared error by up to 26%. Second, a unified autoregressive hybrid block replaces the original chain of 24 specialised modules, eliminating overfitting to specific lead times. Third, the physical core is integrated with two state-of-the-art neural backbones, resulting in PI-PredFormer and PI-IAM4VP. Evaluation on the WeatherBench South Pacific subset from 2000 to 2004 shows that these hybrids reduce root mean squared error at 1-12 h lead times by 8-22% compared to purely neural counterparts, while better preserving physical consistency. These results demonstrate that incremental refinement of hybrid components offers a practical route toward more accurate and efficient short-range weather forecasting.

cs.LG↗

arXiv2026

Ensemble Distributionally Robust Bayesian Optimisation with Continuous Context

We study Bayesian Optimisation (BO) in settings where the objective function is influenced by uncontrollable environmental contexts governed by an unknown probability distribution. In practice, the contextual distribution must be estimated from empirical data, a process that inherently introduces distributional mismatch, producing sub-optimal results. While Distributionally Robust Optimisation (DRO) provides a framework to mitigate these risks, existing robust BO methods frequently suffer from high computational complexity, rely on discretisation of continuous context spaces, or impose restrictive assumptions on the structure of the ambiguity set. To overcome these limitations, we propose Ensemble Distributionally Robust Bayesian Optimisation (EDRBO). Our framework leverages the expressive power of ensemble surrogate models to approximate the black-box function while simultaneously accounting for contextual uncertainty. By utilising Wasserstein ball as ambiguity sets, EDRBO provides a robustified acquisition function that remains computationally tractable and natively handles continuous context spaces. We establish a rigorous theoretical foundation for our approach by proving sublinear cumulative regret guarantees of order $\mathcal{O}(\gamma_T \sqrt{T})$, where $\gamma_T$ represents the maximum information gain within the ensemble. Finally, we provide extensive empirical evaluations that corroborate our theory and demonstrate the state-of-the-art performance of EDRBO.

cs.LG↗

arXiv2026

RL-ABC: Reinforcement Learning for Accelerator Beamline Control

Particle accelerator beamline optimization is a high-dimensional control problem traditionally requiring significant expert intervention. We present RLABC (Reinforcement Learning for Accelerator Beamline Control), an open-source Python framework that automatically transforms standard Elegant beamline configurations into reinforcement learning environments. RLABC integrates with the widely-used Elegant beam dynamics simulation code via SDDS-based interfaces, enabling researchers to apply modern RL algorithms to beamline optimization with minimal RL-specific development. The main contribution is a general methodology for formulating beamline tuning as a Markov decision process: RLABC automatically preprocesses lattice files to insert diagnostic watch points before each tunable element, constructs a 57-dimensional state representation from beam statistics, covariance information, and aperture constraints, and provides a configurable reward function for transmission optimization. The framework supports multiple RL algorithms through Stable-Baselines3 compatibility and implements stage learning strategies for improved training efficiency. Validation on a test beamline derived from the VEPP-5 injection complex (37 control parameters across 11 quadrupoles and 4 dipoles) demonstrates that the framework successfully enables RL-based optimization, with a Deep Deterministic Policy Gradient agent achieving 70.3\% particle transmission -- performance matching established methods such as differential evolution. The framework's stage learning capability allows decomposition of complex optimization problems into manageable subproblems, improving training efficiency. The complete framework, including configuration files and example notebooks, is available as open-source software to facilitate adoption and further research.

cs.LG↗

arXiv2025

The Ratan Active Region Patches (RARPs) Database: A New Database of Solar Active Region Radio Signatures from the RATAN-600 Telescope

Solar flares and coronal mass ejections, originating from solar active regions (ARs), are the primary drivers of space weather and can disrupt technological systems. Forecasting efforts heavily rely on photospheric magnetic field data from the Space-weather HMI Active Region Patch (SHARPs) data products. However, the crucial energy release occurs higher in the solar corona. Radio observations from instruments like the RATAN-600 telescope directly probe this region, but their scientific use has been hindered by a lack of standardized and accessible data products. To address this gap, we have developed the Ratan Active Region Patches (RARPs) database, a new public resource of multi-frequency radio spectra for solar ARs. Generated using RATANSunPy software, RARPs provides the first standardized radio counterpart to magnetic field archives. The database contains over 160,000 calibrated AR observations from 2009 to 2025, each including 3-18 GHz spectra and rich metadata. We demonstrate the scientific utility of this database by using machine learning to forecast solar flares. The radio spectra are first compressed into low-dimensional embedded features using an autoencoder, which are then used as predictors in baseline logistic regression classifiers. We compare the predictive power of these embedded RARPs features with that of the 18 SHARPs magnetic field parameters provided in the SHARPs data product headers. Our results show that while SHARPs data provides superior flare discrimination, the radio signatures in RARPs possess clear predictive potential and, for M-class and above flares, yield lower Brier Scores and positive Brier Skill Scores relative to SHARPs, indicating more accurate probabilistic forecasts for these events. This establishes radio data as a valuable and complementary information source.

astro-ph.SR↗

arXiv2025

Reinforcement Learning for Accelerator Beamline Control: a simulation-based approach

Particle accelerators play a pivotal role in advancing scientific research, yet optimizing beamline configurations to maximize particle transmission remains a labor-intensive task requiring expert intervention. In this work, we introduce RLABC (Reinforcement Learning for Accelerator Beamline Control), a Python-based library that reframes beamline optimization as a reinforcement learning (RL) problem. Leveraging the Elegant simulation framework, RLABC automates the creation of an RL environment from standard lattice and element input files, enabling sequential tuning of magnets to minimize particle losses. We define a comprehensive state representation capturing beam statistics, actions for adjusting magnet parameters, and a reward function focused on transmission efficiency. Employing the Deep Deterministic Policy Gradient (DDPG) algorithm, we demonstrate RLABC's efficacy on two beamlines, achieving transmission rates of 94% and 91%, comparable to expert manual optimizations. This approach bridges accelerator physics and machine learning, offering a versatile tool for physicists and RL researchers alike to streamline beamline tuning.

physics.acc-ph↗

arXiv2025

Distilling Normalizing Flows

Explicit density learners are becoming an increasingly popular technique for generative models because of their ability to better model probability distributions. They have advantages over Generative Adversarial Networks due to their ability to perform density estimation and having exact latent-variable inference. This has many advantages, including: being able to simply interpolate, calculate sample likelihood, and analyze the probability distribution. The downside of these models is that they are often more difficult to train and have lower sampling quality. Normalizing flows are explicit density models, that use composable bijective functions to turn an intractable probability function into a tractable one. In this work, we present novel knowledge distillation techniques to increase sampling quality and density estimation of smaller student normalizing flows. We seek to study the capacity of knowledge distillation in Compositional Normalizing Flows to understand the benefits and weaknesses provided by these architectures. Normalizing flows have unique properties that allow for a non-traditional forms of knowledge transfer, where we can transfer that knowledge within intermediate layers. We find that through this distillation, we can make students significantly smaller while making substantial performance gains over a non-distilled student. With smaller models there is a proportionally increased throughput as this is dependent upon the number of bijectors, and thus parameters, in the network.

cs.LG↗

arXiv2025

Learning to Hear Broken Motors: Signature-Guided Data Augmentation for Induction-Motor Diagnostics

The application of machine learning (ML) algorithms in the intelligent diagnosis of three-phase engines has the potential to significantly enhance diagnostic performance and accuracy. Traditional methods largely rely on signature analysis, which, despite being a standard practice, can benefit from the integration of advanced ML techniques. In our study, we innovate by combining ML algorithms with a novel unsupervised anomaly generation methodology that takes into account the engine physics model. We propose Signature-Guided Data Augmentation (SGDA), an unsupervised framework that synthesizes physically plausible faults directly in the frequency domain of healthy current signals. Guided by Motor Current Signature Analysis, SGDA creates diverse and realistic anomalies without resorting to computationally intensive simulations. This hybrid approach leverages the strengths of both supervised ML and unsupervised signature analysis, achieving superior diagnostic accuracy and reliability along with wide industrial application. The findings highlight the potential of our approach to contribute significantly to the field of engine diagnostics, offering a robust and efficient solution for real-world applications.

cs.LG↗

arXiv2025

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

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.

cs.NE↗

arXiv2025

Optimisation of the Accelerator Control by Reinforcement Learning: A Simulation-Based Approach

Optimizing accelerator control is a critical challenge in experimental particle physics, requiring significant manual effort and resource expenditure. Traditional tuning methods are often time-consuming and reliant on expert input, highlighting the need for more efficient approaches. This study aims to create a simulation-based framework integrated with Reinforcement Learning (RL) to address these challenges. Using \texttt{Elegant} as the simulation backend, we developed a Python wrapper that simplifies the interaction between RL algorithms and accelerator simulations, enabling seamless input management, simulation execution, and output analysis. The proposed RL framework acts as a co-pilot for physicists, offering intelligent suggestions to enhance beamline performance, reduce tuning time, and improve operational efficiency. As a proof of concept, we demonstrate the application of our RL approach to an accelerator control problem and highlight the improvements in efficiency and performance achieved through our methodology. We discuss how the integration of simulation tools with a Python-based RL framework provides a powerful resource for the accelerator physics community, showcasing the potential of machine learning in optimizing complex physical systems.

physics.acc-ph↗

arXiv2025

Engineering Point Defects in MoS2 for Tailored Material Properties using Large Language Models

The tunability of physical properties in transition metal dichalcogenides (TMDCs) through point defect engineering offers significant potential for the development of next-generation optoelectronic and high-tech applications. Building upon prior work on machine learning-driven material design, this study focuses on the systematic introduction and manipulation of point defects in MoS2 to tailor their properties. Leveraging a comprehensive dataset generated via density functional theory (DFT) calculations, we explore the effects of various defect types and concentrations on the mate rial characteristics of TMDCs. Our methodology integrates the use of pre-trained large language models to generate defect configurations, enabling efficient predictions of defect-induced property modifications. This research differs from traditional methods of material generation and discovery by utilizing the latest advances in transformer model architecture, which have proven to be efficient and accurate discrete predictors. In contrast to high-throughput methods where configurations are generated randomly and then screened based on their physical properties, our approach not only enhances the understanding of defect-property relationships in TMDCs but also provides a robust framework for designing materials with bespoke properties. This facilitates the advancement of materials science and technology.

cond-mat.mtrl-sci↗

arXiv2024

Intelligent Algorithms For Signature Diagnostics Of Three-Phase Motors

The application of machine learning (ML) algorithms in the intelligent diagnosis of three-phase engines has the potential to significantly enhance diagnostic performance and accuracy. Traditional methods largely rely on signature analysis, which, despite being a standard practice, can benefit from the integration of advanced ML techniques. In our study, we innovate by combining state of the art algorithms with a novel unsupervised anomaly generation methodology that takes into account physics model of the engine. This hybrid approach leverages the strengths of both supervised ML and unsupervised signature analysis, achieving superior diagnostic accuracy and reliability along with a wide industrial application. Our experimental results demonstrate that this method significantly outperforms existing ML and non-ML state-of-the-art approaches while retaining the practical advantages of an unsupervised methodology. The findings highlight the potential of our approach to significantly contribute to the field of engine diagnostics, offering a robust and efficient solution for real-world applications.

eess.SP↗

arXiv2024

Global Optimisation of Black-Box Functions with Generative Models in the Wasserstein Space

We propose a new uncertainty estimator for gradient-free optimisation of black-box simulators using deep generative surrogate models. Optimisation of these simulators is especially challenging for stochastic simulators and higher dimensions. To address these issues, we utilise a deep generative surrogate approach to model the black box response for the entire parameter space. We then leverage this knowledge to estimate the proposed uncertainty based on the Wasserstein distance - the Wasserstein uncertainty. This approach is employed in a posterior agnostic gradient-free optimisation algorithm that minimises regret over the entire parameter space. A series of tests were conducted to demonstrate that our method is more robust to the shape of both the black box function and the stochastic response of the black box than state-of-the-art methods, such as efficient global optimisation with a deep Gaussian process surrogate.

cs.LG↗

arXiv2023

The LHCb ultra-fast simulation option, Lamarr: design and validation

Detailed detector simulation is the major consumer of CPU resources at LHCb, having used more than 90% of the total computing budget during Run 2 of the Large Hadron Collider at CERN. As data is collected by the upgraded LHCb detector during Run 3 of the LHC, larger requests for simulated data samples are necessary, and will far exceed the pledged resources of the experiment, even with existing fast simulation options. An evolution of technologies and techniques to produce simulated samples is mandatory to meet the upcoming needs of analysis to interpret signal versus background and measure efficiencies. In this context, we propose Lamarr, a Gaudi-based framework designed to offer the fastest solution for the simulation of the LHCb detector. Lamarr consists of a pipeline of modules parameterizing both the detector response and the reconstruction algorithms of the LHCb experiment. Most of the parameterizations are made of Deep Generative Models and Gradient Boosted Decision Trees trained on simulated samples or alternatively, where possible, on real data. Embedding Lamarr in the general LHCb Gauss Simulation framework allows combining its execution with any of the available generators in a seamless way. Lamarr has been validated by comparing key reconstructed quantities with Detailed Simulation. Good agreement of the simulated distributions is obtained with two-order-of-magnitude speed-up of the simulation phase.

hep-ex↗

arXiv2023

Symbolic expression generation via Variational Auto-Encoder

There are many problems in physics, biology, and other natural sciences in which symbolic regression can provide valuable insights and discover new laws of nature. A widespread Deep Neural Networks do not provide interpretable solutions. Meanwhile, symbolic expressions give us a clear relation between observations and the target variable. However, at the moment, there is no dominant solution for the symbolic regression task, and we aim to reduce this gap with our algorithm. In this work, we propose a novel deep learning framework for symbolic expression generation via variational autoencoder (VAE). In a nutshell, we suggest using a VAE to generate mathematical expressions, and our training strategy forces generated formulas to fit a given dataset. Our framework allows encoding apriori knowledge of the formulas into fast-check predicates that speed up the optimization process. We compare our method to modern symbolic regression benchmarks and show that our method outperforms the competitors under noisy conditions. The recovery rate of SEGVAE is 65% on the Ngyuen dataset with a noise level of 10%, which is better than the previously reported SOTA by 20%. We demonstrate that this value depends on the dataset and can be even higher.

cs.LG↗

arXiv2022

Robust Neural Particle Identification Models

The volume of data processed by the Large Hadron Collider experiments demands sophisticated selection rules typically based on machine learning algorithms. One of the shortcomings of these approaches is their profound sensitivity to the biases in training samples. In the case of particle identification (PID), this might lead to degradation of the efficiency for some decays not present in the training dataset due to differences in input kinematic distributions. In this talk, we propose a method based on the Common Specific Decomposition that takes into account individual decays and possible misshapes in the training data by disentangling common and decay specific components of the input feature set. We show that the proposed approach reduces the rate of efficiency degradation for the PID algorithms for the decays reconstructed in the LHCb detector.

hep-ex↗

arXiv2022

New UTfit Analysis of the Unitarity Triangle in the Cabibbo-Kobayashi-Maskawa scheme

Flavour mixing and CP violation as measured in weak decays and mixing of neutral mesons are a fundamental tool to test the Standard Model (SM) and to search for new physics. New analyses performed at the LHC experiment open an unprecedented insight into the Cabibbo-Kobayashi-Maskawa (CKM) metrology and new evidence for rare decays. Important progress has also been achieved in theoretical calculations of several hadronic quantities with a remarkable reduction of the uncertainties. This improvement is essential since previous studies of the Unitarity Triangle did show that possible contributions from new physics, if any, must be tiny and could easily be hidden by theoretical and experimental errors. Thanks to the experimental and theoretical advances, the CKM picture provides very precise SM predictions through global analyses. We present here the results of the latest global SM analysis performed by the UTfit collaboration including all the most updated inputs from experiments, lattice QCD and phenomenological calculations.

hep-ph↗