Deep Learning for and in Physics - My Papers
Deep learning for (anti)neutrino-nuclei scattering
The project’s goal is to develop deep learning tools for neutrino interactions.
Jose L. Bonilla, Krzysztof M. Graczyk, Artur M. Ankowski, Rwik Dharmapal Banerjee, Beata E. Kowal, Hemant Prasad, Jan T. Sobczyk, Generative adversarial neural networks for simulating neutrino interactions, arxiv:2502.20244
Krzysztof M. Graczyk, Beata Kowal, Artur M. Ankowski, Rwik D. Banerjee, Jose Luis Bonilla, Hemant Prasad, Jan T. Sobczyk, Electron-nucleus cross sections from transfer learning, arxiv:2408.09936
Beata Kowal, Krzysztof M. Graczyk, Artur M. Ankowski, Rwik D. Banerjee, Hemant Prasad, Jan T. Sobczyk, Empirical fits to inclusive electron-carbon scattering data obtained by deep-learning methods, Phys. Rev. C 110, 025501
- The fits are available at repository
Physics Informed Neural Networks (PINNs)
The goal of the project is to propose Bayesian framework for PINNs
- Krzysztof M. Graczyk, Kornel Witkowski, Bayesian Reasoning for Physics Informed Neural Networks arxiv:2312.13222
- see also »»
Deep learning for porous media
The goal of the project is to develop the deep learning tools with abilities to predict the macroscopic and microscopic properties of the fluid flow and diffusion phenomena in porous materials
- Krzysztof M. Graczyk, Dawid Strzelczyk and Maciej Matyka, Deep learning for diffusion in porous media, Sci Rep 13, 9769 (2023)
- Krzysztof M. Graczyk, Maciej Matyka, Predicting Porosity, Permeability, and Tortuosity of Porous Media from Images by Deep Learning, Sci Rep 10, 21488 (2020)
Uncertainties in Deep Learning Systems Counting Microbiological Objects
The project aimed to develop the methods to estimate the uncertainties in density map model (given by $U^2$-Net) predictions.
The project supported by MOZART Grant WCA, conducted in NeuroSys
- Krzysztof M. Graczyk, Jarosław Pawlowski, Sylwia Majchrowska, Tomasz Golan, Self-Normalized Density Map (SNDM) for Counting Microbiological Obejcts, Sci Rep 12, 10583 (2022)
Bayesian neural networks for nuclean electrowak structure studies
Bayesian neural networks were adapted to analyze:
- elastic electron-proton and electron-neutron scattering data
- quasielastic neutrino-deuteron scattering data
The Bayesian framework allowed us to:
- Extract two-photon exchange correction
- Obtain model-independent electromagnetic form factors of the nucleon
- Estimate the proton radius
- Nucleon axial form factor
To conduct the project, the C++ library for Bayesian Neural Networks was developed (together with Cezary Juszczak)
Luis Alvarez-Ruso, Krzysztof M. Graczyk, Eduardo Saul-Sala, Nucleon axial form factor from a Bayesian neural-network analysis of neutrino-scattering data, Phys. Rev. C99, 025204 (2019)
Krzysztof M. Graczyk and Cezary Juszczak, Zemach moments of proton from Bayesian inference, Phys. Rev. C91, 045205 (2015)
Krzysztof M. Graczyk and Cezary Juszczak, Applications of Neural Networks in Hadron Physics, J.Phys. G42 (2015) 3, 034019
Krzysztof M. Graczyk and Cezary Juszczak, Proton Radius from Bayesian Inference, Phys. Rev. C90, 054334 (2014)
Krzysztof M. Graczyk, Comparison of Neural Network and Hadronic Model Predictions of Two-Photon Exchange Effect, Phys. Rev. C88, 065205 (2013)
Krzysztof M. Graczyk, Two-photon exchange effect studied with neural networks, Phys. Rev. C84, 034314 (2011)
Krzysztof M. Graczyk, Piotr Płoński, Robert Sulej, Neural Network Parameterizations of Electromagnetic Nucleon Form Factors, JHEP (2010) 053