Применение генеративных моделей для физических экспериментов тема диссертации и автореферата по ВАК РФ 00.00.00, кандидат наук Рогачев Александр Игоревич

• Рогачев Александр Игоревич
• кандидат наук
• 2025

Оглавление диссертации кандидат наук Рогачев Александр Игоревич

Contents

Chapter 1 Introduction

1.1 New Physics and the LHCb Experiment in Search of It

1.2 Machine Learning

1.3 My Contribution

Chapter 2 Machine Learning

2.1 A Brief History of Artificial Intelligence and Machine Learning

2.2 Distinction between Traditional Algorithms and ML

2.3 Machine Learning Formalism

2.3.1 Basic Concepts

2.3.2 Learning types

2.3.3 Task types

2.3.4 No Free Lunch Theorem

2.4 Measuring Model Quality

2.4.1 Quality Metrics vs. Loss Functions

2.4.2 Metrics Examples

2.4.3 Hyperparameter tuning

2.4.4 Cross-validation

2.5 Deep Learning

2.5.1 The Basic Architecture of Neural Networks

2.5.2 Activation Functions

2.5.3 Deep Neural Networks

2.5.4 Convolutional Neural Networks

2.5.5 Optimization

2.5.6 Training Pipeline

2.5.7 Conclusion

2.6 Generative models

2.6.1 Basic Principles

2.6.2 Generative Adversarial Networks

2.6.3 Wasserstein GAN

2.6.4 Other models

2.7 Conclusion

Chapter 3 Machine Learning in High-Energy Physics

3.1 Computational Challenges in Simulations

3.2 Historical Context

3.3 Generative Models for Fast Simulation

3.4 Evaluating Generative Models in HEP

3.4.1 Evaluation criterias

3.4.2 Integral Probability Metrics

3.4.3 NN-based metrics

3.4.4 Manifold estimation

3.4.5 Metrics selection

3.5 Conclusion

Chapter 4 The LHCb experiment

4.1 The Large Hadron Collider

4.2 The LHCb Detector

4.3 Calorimeters

4.3.1 Purpose of ECAL and Fast Simulation

4.3.2 Construction

Chapter 5 GAN for Calorimeter Response

5.1 Background

5.2 Dataset

5.3 Quality Evaluation

5.4 CaloGAN Improvement

5.4.1 Self-Attention

5.4.2 Hinge Loss

5.4.3 Spectral Normalization

5.5 Experiments

5.5.1 Number of Clusters for Evaluation

5.5.2 Results

5.6 Conclusion

Chapter 6 Auxiliary Regressors

6.1 Auxiliary Regressor for GANs

6.2 Method

6.3 Detailed Quality Evaluation

6.4 General Experiments

6.5 Controlling Quality

6.5.1 Auxiliary Loss Weight

6.5.2 Auxiliary Regressor's Quality Impact

6.5.3 Training Setting

6.6 Conclusion

Chapter 7 Soft Spectral Normalization

7.1 Capacity Reduction Case

7.1.1 Fast Calorimeter Simulation Challenge

7.1.2 Low Capacity

7.2 Lipschitz Networks

7.3 Lipschitz Regularization

7.4 Method

7.5 General Experiments

7.6 Hyperparameters Relationship

7.7 Conclusion

Chapter 8 Conclusion

Bibliography

Appendix A Examples of generation

Appendix B Fast Calorimeter Simulation Challenge

Appendix C Russian Translation of the Dissertation

Введение диссертации (часть автореферата) на тему «Применение генеративных моделей для физических экспериментов»

Chapter 1

Introduction

1.1 New Physics and the LHCb Experiment in Search of It

The quest for new physics beyond the Standard Model (SM) is one of the most compelling endeavors in modern high-energy physics. The Large Hadron Collider beauty (LHCb) experiment at CERN plays a pivotal role in this pursuit. The LHCb detector is specifically designed to study particles containing b and c quarks, which are produced in proton-proton collisions at the LHC. By examining the properties and interactions of these particles, the LHCb experiment aims to uncover phenomena that cannot be explained by the Standard Model alone [1].

One of the primary objectives of the LHCb experiment is to investigate the matter-antimatter asymmetry in the universe, known as CP violation. Precise measurements of CP violation in the decays of b and c hadrons can provide crucial insights into the underlying mechanisms that led to the dominance of matter over antimatter in the universe [2]. Additionally, the LHCb experiment searches for rare decays and exotic particles that could signal the presence of new physics. For instance, the observation of them can provide stringent tests of the SM and potential signs of new physics [3]. Finding and studying rare decays means pushing the frontiers of experiment design and data analysis. During its lifetime, the LHC provides an unprecedented, but still finite number of collision events. The rarer the process, the more is required from the experimental hardware and software for the measurement to be statistically significant. It is fundamental to develop selection criteria, particle identification, and background parametrization to enable the discovery of rare decays and place even more stringent limits on supersymmetry and other new physics models.

The LHCb detector is equipped with advanced sub-detectors, including vertex detectors, tracking systems, and calorimeters, which enable it to perform high-precision measurements. Among these, the electromagnetic calorimeter (ECAL) plays a critical role in identifying and measuring the energy of photons and electrons. Accurate simulation of the ECAL response is essential for the success of the LHCb experiment, as it directly impacts the reconstruction and analysis of particle interactions [4].

The complexities of processing experimental data extend beyond mere data col-

lection; simulation plays a vital role in the design and construction of experiments, as well as in the development of algorithms for data analysis. However, this simulation process incurs substantial computational costs. During Run 2, Monte Carlo generation alone accounted for approximately 75% of CPU time utilized within the LHCb GRID. With the planned escalation in luminosity, these computational demands are projected to become unsustainable, necessitating innovative solutions.

Data-driven methods present a promising alternative to traditional simulation techniques for evaluating and refining ECAL algorithms. While these methods leverage calibration samples—comprising charged tracks of various species selected independently of the ECAL response—they also introduce complexities, such as potential biases from selection processes and background interference. The existing accuracy of ECAL responses derived from these methods is often inadequate for most analyses, highlighting the need for improvement.

The interdisciplinary nature of this research, which merges concepts from computer science with high-energy physics, emphasizes a novel approach to addressing some of the field's most pressing challenges. The methodologies developed not only advance the state of fast simulation but also contribute to the overarching goal of uncovering new physics phenomena and general deep learning.

1.2 Machine Learning

Generative models, as an integral component of machine learning, have the transformative capability to drive advancements in various scientific disciplines, including high-energy physics. These models, which include Variational Autoencoders (VAEs) [5] and Generative Adversarial Networks (GANs) [6], have garnered significant attention for their ability to learn complex data distributions and generate new data instances that mimic the original data. This property is particularly beneficial in high-energy physics, where the generation of synthetic data can play a pivotal role in experiment simulations, data augmentation, and anomaly detection.

High-energy physics experiments, such as those conducted at the Large Hadron Collider (LHC), generate vast amounts of data. The analysis of this data is crucial for understanding the fundamental particles and forces that govern the universe. However, the sheer volume and complexity of the data pose significant challenges. Traditional simulation methods, while effective, are computationally intensive and time-consuming. Generative models offer a promising solution by enabling the efficient generation of synthetic data that accurately reflects the underlying physical processes.

The application of generative models in high-energy physics has already shown promising results. For instance, GANs have been successfully employed to simulate the energy deposits of particles in calorimeters, a task that is both critical and computationally demanding. These models have demonstrated the ability to produce high-fidelity simulations significantly faster than traditional methods, thereby accelerating the research pipeline.

Furthermore, generative models have the potential to uncover new insights into the data generated by high-energy physics experiments. By learning the complex distributions of this data, these models can aid in identifying rare events or anoma-

lies that may signal new physics beyond the current understanding. This capability makes generative models a powerful tool for exploration and discovery in the field.

However, the application of generative models in high-energy physics also poses unique challenges. The accurate simulation of physical processes requires models that can capture the intricate details and constraints of the data. Additionally, the evaluation of model performance in this context is non-trivial, as it involves not only the fidelity of the generated data but also its physical plausibility.

Currently, generative models represent a cutting-edge approach that can significantly enhance the capabilities of high-energy physics research. By enabling the efficient generation of synthetic data and offering new avenues for exploration, these models hold the promise of accelerating discoveries and deepening our understanding of the universe. The continued development and refinement of generative models, tailored to the specific needs and challenges of high-energy physics, will undoubtedly be a fruitful area of research in the years to come.

1.3 My Contribution

This thesis represents a significant advancement in the domain of fast simulation within high-energy physics (HEP), focusing on the development and critical evaluation of generative models, particularly within the context of the LHCb experiment and its calorimeter response simulation. By addressing key challenges in simulation accuracy and efficiency, this work contributes to improving computational techniques used in experimental physics.

The theoretical significance of this research is underscored by a comprehensive exploration of the LHCb experiment and its calorimeter response simulation. The study introduces improvements to the CaloGAN model, including self-attention mechanisms, hinge loss, and spectral normalization, demonstrating the potential of Generative Adversarial Networks (GANs) in enhancing fast simulations. Furthermore, the investigation of Lipschitz networks emphasizes the importance of theoretical research in understanding model stability and training dynamics.

Practically, this research has led to the development of methods that significantly improve the accuracy and efficiency of fast simulations, which are crucial for the LHCb experiment's ability to analyze large datasets and search for new physics phenomena. The introduction of auxiliary regressors and soft spectral normalization methods enhances the reproduction quality of key metrics and mitigates capacity reduction in GANs. These advancements have applications beyond HEP, extending to other domains where generative models are utilized.

The thesis is structured as follows:

• Chapter 2 introduces the LHCb experiment, providing an overview of its goals, design, and operational aspects. It highlights its role in studying CP violation and rare hadron decays and discusses challenges in data collection and analysis, motivating the need for advanced computational techniques such as machine learning.

• Chapter 3 provides a general overview of machine learning, covering fundamental concepts, algorithms, and the evolution of ML techniques.

• Chapter 4 explores the application of machine learning in HEP, reviewing previous work, discussing challenges and opportunities, and setting the stage for the integration of generative models in this domain.

• Chapter 5 presents the application of GANs for calorimeter response simulation, discussing existing approaches, model architectures, and the evaluation of generative models.

• Chapter 6 introduces the auxiliary regressor method, explaining its role in improving generative model performance and experimental results.

• Chapter 7 discusses soft spectral normalization and its impact on training stability and model capacity, providing insights into balancing performance and stability.

My individual contributions include:

Generative Adversarial Networks for Calorimeter Response I contributed to the continuous development of the CaloGAN model, proposing architectural improvements and refining quality metrics to enhance simulation accuracy. These advancements are detailed in Chapter 5 and [7, 8].

Auxiliary Regressors for Improved Generation Quality I proposed and implemented the Auxiliary Regressor approach to extend the capabilities of the GAN discriminator. This method improves the reproduction of key physics metrics critical for evaluating generated data quality. The details of this approach are discussed in Chapter 6 and [9, 10].

Soft Spectral Normalization for Training Stability To address training stability challenges introduced by auxiliary regression, I developed Soft Spectral Normalization, an adaptation of spectral normalization from [11]. This method stabilizes training while preserving model flexibility. Details are provided in Chapter 7 and [12, 13].

Заключение диссертации по теме «Другие cпециальности», Рогачев Александр Игоревич

Conclusion

This thesis advances the field of fast simulation in high-energy physics through the development and evaluation of generative models, specifically focusing on the LHCb experiment and calorimeter response simulation. The work spans several essential areas, each contributing to the overarching goal of improving simulation accuracy and efficiency.

Chapters 2 and 3 provided a comprehensive overview of the historical context and the role of simulations in HEP. It discussed the evolution of generative models for fast simulation and introduces various evaluation criteria, including Integral Probability Metrics and NN-based metrics. The chapter concludes with a discussion on metrics selection, setting the stage for the subsequent detailed evaluations.

Chapter 4 delved into the specifics of the LHCb experiment, detailing the Large Hadron Collider and the LHCb Detector. It emphasizes the importance of the ECAL and fast simulation, describing their construction and purpose. This foun-dational knowledge is crucial for understanding the challenges and requirements of calorimeter response simulation.

Chapter 5 focused on the application of Generative Adversarial Networks (GANs) for simulating calorimeter responses. It introduced the dataset used and the quality evaluation metrics. The chapter mainly discussed improvements to the CaloGAN model, including the incorporation of self-attention, hinge loss, and spectral normalization. Experimental results demonstrate the effectiveness of these enhancements, providing a robust framework for future work.

Chapter 6 explored the use of auxiliary regressors to further improve GAN performance. It outlined the method and provided a detailed quality evaluation. The chapter also discussed comparative experiments and the impact of auxiliary loss weight on the generation quality. The relationship between training settings and model performance was examined, offering insights into optimizing the training process.

Chapter 7 introduced the soft spectral normalization method, addressing the issue of capacity reduction in GANs. It discussed the theoretical foundations of Lipschitz networks and regularization, and presented the methodology for implementing soft spectral normalization. General experiments validate the method's efficacy, and the relationship between hyperparameters was explored to guide future optimizations.

Overall, the developed methods and approaches represent a significant interdisciplinary contribution to both computer science and physics. By improving the accuracy and efficiency of fast simulations, these advancements enhance the capability of current HEP experiments to search for new physics phenomena. The work lays a solid foundation for more sensitive and sophisticated data processing in future experiments, pushing the boundaries of what is possible in high-energy physics research.

The author of the thesis is a main author of the proposed auxiliary regression approach, implemented and set all the experiments in [9, 10]; proposed soft margin spectral normalization, implemented and set all the experiments in [12, 13]; the main author and coauthor of the rest related application studies, suggested architectural improvements, implemented and set all the experiments in [7, 8].

The results of the research were partially presented during multiple conferences with same report titles as papers, such as:

• 25th International Conference on Computing in High Energy and Nuclear Physics (CHEP 2021), 17-21 May 2021, online, where [7] was discussed;

• [8] was presented during 15th Pisa Meeting on Advanced Detectors, Isola d'Elba, Italy, May 22-28, 2022;

• [9] was introduced on 20th International Workshop on Advanced Computing and Analysis Techniques in Physics Research, 29 November-3 December 2021, online;

• [10] was shown during 26th International Conference on Computing in High Energy and Nuclear Physics (CHEP 2023),Norfolk, USA, May 8-12, 2023;

• [12] won the prize during AIJ 2023, Moscow, Russia, 22-24 November 2023, as the best paper

• [13] was recently published

Thus, the following essential frontiers were advanced:

Fast Simulation of the ECAL Response The quality of the generation of the electromagnetic calorimeter's response was consistently improved through the proposed methods from chapters 5, 6 and 7. These result provide a reasonable candidate for the production inference, developing the cutting edge technologies for the needs of LHCb collaboration.

Generative Adversarial Networks The introduced auxiliary method can be applied not only for the LHCb case, thus it extends the general abilities of GANs, providing an approach to guide the network. The proposed evaluation criteria used in the thesis may be generalized for the other domains as well.

Deep Learning and Neural Networks The soft spectral normalization was analyzed in the context of the generative adversarial networks, but may be extended further as it has no limits related to GANs only. As aux-regressors, this introduction opens new ways for the deeper investigation and research, as the Lipschitz properties of the networks are still needed to be considered and developed.

Overall The methods and approaches developed in this thesis represent a significant interdisciplinary effort bridging machine learning and physics. These advancements enhance the state-of-the-art in high-energy physics simulations, particularly in the context of the LHCb experiment and calorimeter response. By improving the accuracy and efficiency of generative models and their evaluation, these methods significantly boost the performance of current experiments in the search for new physics phenomena. Furthermore, they lay a robust foundation for even more sensitive and sophisticated data processing techniques in future experiments, thereby pushing the boundaries of what can be achieved in high-energy physics research.

Список литературы диссертационного исследования кандидат наук Рогачев Александр Игоревич, 2025 год

Bibliography

[1] A. A. Alves Jr. et al. "The LHCb detector at the LHC." In: JINST 3 (2008), S08005. doi: 10.1088/1748-0221/3/08/S08005.

[2] A. J. Bevan et al. "The Physics of the B Factories." In: The European Physical Journal C 74.11 (Nov. 2014), p. 3026. issn: 1434-6052. doi: 10.1140/epjc/ s10052-014-3026-9. url: https://doi.org/10.1140/epjc/s10052-014-3026-9.

[3] Roel Aaij et al. "First observation of forward Z ^ bb production in pp collisions at ^s = 8 TeV." In: Physics Letters B 776 (2018), pp. 430-439.

[4] CERN (Meyrin) LHCb Collaboration. LHCb Upgrade GPU High Level Trigger Technical Design Report. Tech. rep. CERN-LHCC-2020-006. LHCB-TDR-021. Geneva: CERN, May 2020. url: https://cds.cern.ch/record/ 2717938.

[5] Diederik P Kingma and Jimmy Ba. "Adam: A method for stochastic optimization." In: arXiv preprint arXiv:1412.6980 (2014).

[6] Ian Goodfellow et al. "Generative adversarial nets." In: Advances in neural information processing systems. 2014, pp. 2672-2680.

[7] Fedor Ratnikov and Alexander Rogachev. "Fast simulation of the electromagnetic calorimeter response using Self-Attention Generative Adversarial Networks." In: EPJ Web of Conferences. Vol. 251. EDP Sciences. 2021, p. 03043.

[8] Fedor Ratnikov et al. "A full detector description using neural network driven simulation." In: Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 1046 (2023), p. 167591.

[9] Alexander Rogachev and Fedor Ratnikov. "GAN with an auxiliary regres-sor for the fast simulation of the electromagnetic calorimeter response." In: Journal of Physics: Conference Series. Vol. 2438. 1. IOP Publishing. 2023, p. 012086.

[10] Alexander Rogachev and Fedor Ratnikov. "Controlling Quality for a Physics-Driven Generative Models and Auxiliary Regression Approach." In: EPJ Web of Conferences. Vol. 295. EDP Sciences. 2024, p. 09007.

[11] Takeru Miyato et al. Spectral Normalization for Generative Adversarial Networks. 2018. arXiv: 1802.05957 [cs.LG].

[12] EA Egorov and Aleksandr Igorevich Rogachev. "Adaptive spectral normalization for generative models." In: Doklady Mathematics. Vol. 108. Suppl 2. Springer. 2023, S205-S214.

[13] Alexander Rogachev and Fedor Ratnikov. "Soft Margin Spectral Normalization for GANs." In: Computing and Software for Big Science 8.1 (2024), p. 12.

[14] John McCarthy et al. "A proposal for the dartmouth summer research project on artificial intelligence, august 31, 1955." In: AI magazine 27.4 (2006), pp. 12-12.

[15] Allen Newell and Herbert Simon. "The logic theory machine-A complex information processing system." In: IRE Transactions on information theory 2.3 (1956), pp. 61-79.

[16] George W Ernst and Allen Newell. "GPS: A case study in generality and problem solving." In: (No Title) (1969).

[17] Bruce G Buchanan. "Expert systems: working systems and the research literature." In: Expert systems 3.1 (1986), pp. 32-50.

[18] Arthur L Samuel. "Some studies in machine learning using the game of checkers." In: IBM Journal of research and development 3.3 (1959), pp. 210-229.

[19] Yann LeCun et al. "Backpropagation applied to handwritten zip code recognition." In: Neural computation 1.4 (1989), pp. 541-551.

[20] Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. "Deep learning." In: nature 521.7553 (2015), pp. 436-444.

[21] Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. "Imagenet classification with deep convolutional neural networks." In: Advances in neural information processing systems. 2012, pp. 1097-1105.

[22] Diederik P Kingma and Max Welling. "Auto-encoding variational bayes." In: arXiv preprint arXiv:1312.6114 (2013).

[23] Jascha Sohl-Dickstein et al. "Deep unsupervised learning using nonequilib-rium thermodynamics." In: International conference on machine learning. PMLR. 2015, pp. 2256-2265.

[24] Jonathan Ho, Ajay Jain, and Pieter Abbeel. "Denoising diffusion probabilistic models." In: Advances in neural information processing systems 33 (2020), pp. 6840-6851.

[25] Martin Luscher. "Computational strategies in lattice QCD." In: Les Houches Summer School: Session 93 (2010), pp. 331-399.

[26] Torbjorn Sjostrand, Stephen Mrenna, and Peter Skands. "PYTHIA 6.4 physics and manual." In: Journal of High Energy Physics 2006.05 (2006), p. 026.

[27] Alexander Radovic et al. "Machine learning at the energy and intensity frontiers of particle physics." In: Nature 560.7716 (2018), p. 41.

[28] Vladimir N Vapnik. "An overview of statistical learning theory." In: IEEE transactions on neural networks 10.5 (1999), pp. 988-999.

[29] Augustin Cauchy. "Méthode générale pour la résolution des systemes d'équations simultanées." In: Comp. Rend. Sci. Paris 25.1847 (1847), pp. 536-538.

[30] Léon Bottou. "Large-scale machine learning with stochastic gradient descent." In: Proceedings of COMPSTAT'2010: 19th International Conference on Computational StatisticsParis France, August 22-27, 2010 Keynote, Invited and Contributed Papers. Springer. 2010, pp. 177-186.

[31] Christopher M Bishop. Pattern recognition and machine learning. springer, 2006.

[32] Anil K Jain, M Narasimha Murty, and Patrick J Flynn. "Data clustering: a review." In: ACM computing surveys (CSUR) 31.3 (1999), pp. 264-323.

[33] Ian T Jolliffe and Jorge Cadima. "Principal component analysis: a review and recent developments." In: Philosophical transactions of the royal society A: Mathematical, Physical and Engineering Sciences 374.2065 (2016), p. 20150202.

[34] Jacob Edwin Crosby. "Searches for New Physics Using Unsupervised Machine Learning for Anomaly Detection at the ATLAS Detector and the Development of Particle Identification Algorithms for the HL-LHC." MA thesis. Oklahoma State U., May 2024.

[35] Jan Kukacka, Vladimir Golkov, and Daniel Cremers. "Regularization for deep learning: A taxonomy." In: arXiv preprint arXiv:1710.10686 (2017).

[36] Jerome Friedman. "The elements of statistical learning: Data mining, inference, and prediction." In: (No Title) (2009).

[37] David H Wolpert and William G Macready. "No free lunch theorems for optimization." In: IEEE transactions on evolutionary computation 1.1 (1997), pp. 67-82.

[38] Tom Fawcett. "An introduction to ROC analysis." In: Pattern recognition letters 27.8 (2006), pp. 861-874.

[39] James Bergstra and Yoshua Bengio. "Random search for hyper-parameter optimization." In: Journal of machine learning research 13.2 (2012).

[40] Jasper Snoek, Hugo Larochelle, and Ryan P Adams. "Practical bayesian optimization of machine learning algorithms." In: Advances in neural information processing systems 25 (2012).

[41] F. Pedregosa et al. "Scikit-learn: Machine Learning in Python." In: Journal of Machine Learning Research 12 (2011), pp. 2825-2830.

[42] Ron Kohavi et al. "A study of cross-validation and bootstrap for accuracy estimation and model selection." In: Ijcai. Vol. 14. 2. Montreal, Canada. 1995, pp.1137-1145.

[43] Ian Goodfellow, Yoshua Bengio, and Aaron Courville. Deep learning. MIT press, 2016.

[44] Jürgen Schmidhuber. "Deep learning in neural networks: An overview." In: Neural networks 61 (2015), pp. 85-117.

[45] Geoffrey E Hinton, Simon Osindero, and Yee-Whye Teh. "A fast learning algorithm for deep belief nets." In: Neural computation 18.7 (2006), pp. 15271554.

[46] Qiming Zhang et al. "Artificial neural networks enabled by nanophotonics." In: Light: Science & Applications 8.1 (2019), p. 42.

[47] Warren S McCulloch and Walter Pitts. "A logical calculus of the ideas immanent in nervous activity." In: The bulletin of mathematical biophysics 5 (1943), pp. 115-133.

[48] Andrew L Maas, Awni Y Hannun, and Andrew Y Ng. "Rectifier nonlinearities improve neural network acoustic models." In: Proc. icml. Vol. 30. 1. 2013, p. 3.

[49] Jun Han and Claudio Moraga. "The influence of the sigmoid function parameters on the speed of backpropagation learning." In: International workshop on artificial neural networks. Springer. 1995, pp. 195-201.

[50] Yann LeCun et al. "Efficient backprop." In: Neural networks: Tricks of the trade. Springer, 2002, pp. 9-50.

[51] Sheng Qian et al. "Adaptive activation functions in convolutional neural networks." In: Neurocomputing 272 (2018), pp. 204-212.

[52] Prajit Ramachandran, Barret Zoph, and Quoc V Le. "Searching for activation functions." In: arXiv preprint arXiv:1710.05941 (2017).

[53] Yoshua Bengio et al. "Learning deep architectures for AI." In: Foundations and trends® in Machine Learning 2.1 (2009), pp. 1-127.

[54] Olivier Delalleau and Yoshua Bengio. "Shallow vs. deep sum-product networks." In: Advances in neural information processing systems 24 (2011).

[55] Ashish Vaswani et al. "Attention is all you need." In: Advances in neural information processing systems 30 (2017).

[56] Yann LeCun et al. "Gradient-based learning applied to document recognition." In: Proceedings of the IEEE 86.11 (1998), pp. 2278-2324.

[57] Andrew M Saxe, James L McClelland, and Surya Ganguli. "Exact solutions to the nonlinear dynamics of learning in deep linear neural networks." In: arXiv preprint arXiv:1312.6120 (2013).

[58] Yann N Dauphin et al. "Identifying and attacking the saddle point problem in high-dimensional non-convex optimization." In: Advances in neural information processing systems. 2014, pp. 2933-2941.

[59] Tijmen Tieleman and Geoffrey Hinton. "Lecture 6.5-rmsprop: Divide the gradient by a running average of its recent magnitude." In: COURSERA: Neural networks for machine learning 4.2 (2012), pp. 26-31.

[60] John Duchi, Elad Hazan, and Yoram Singer. "Adaptive subgradient methods for online learning and stochastic optimization." In: Journal of machine learning research 12.7 (2011).

[61] Matthew D Zeiler. "ADADELTA: an adaptive learning rate method." In: arXiv preprint arXiv:1212.5701 (2012).

[62] Timothy Dozat. "Incorporating nesterov momentum into adam." In: (2016).

[63] Adam Paszke et al. "Pytorch: An imperative style, high-performance deep learning library." In: Advances in neural information processing systems 32 (2019).

[64] CUDA Nvidia. "CUDA toolkit documentation." In: NVIDIA Corp.(NVIDIA), Santa Clara, CA, USA, Tech. Rep 12 (2018).

[65] Xavier Glorot and Yoshua Bengio. "Understanding the difficulty of training deep feedforward neural networks." In: Proceedings of the thirteenth international conference on artificial intelligence and statistics. JMLR Workshop and Conference Proceedings. 2010, pp. 249-256.

[66] Kaiming He et al. "Delving deep into rectifiers: Surpassing human-level performance on imagenet classification." In: Proceedings of the IEEE international conference on computer vision. 2015, pp. 1026-1034.

[67] Noseong Park et al. "Data synthesis based on generative adversarial networks." In: arXiv preprint arXiv:1806.03384 (2018).

[68] Conor Hassan, Robert Salomone, and Kerrie Mengersen. "Deep Generative Models, Synthetic Tabular Data, and Differential Privacy: An Overview and Synthesis." In: arXiv preprint arXiv:2307.15424 (2023).

[69] Garima Agrawal, Amardeep Kaur, and Sowmya Myneni. "A review of generative models in generating synthetic attack data for cybersecurity." In: Electronics 13.2 (2024), p. 322.

[70] Chenyu Zheng, Guoqiang Wu, and Chongxuan Li. "Toward understanding generative data augmentation." In: Advances in Neural Information Processing Systems 36 (2024).

[71] Aleksandra Edwards et al. "Guiding generative language models for data augmentation in few-shot text classification." In: arXiv preprint arXiv:2111.09064 (2021).

[72] Alec Radford, Luke Metz, and Soumith Chintala. "Unsupervised representation learning with deep convolutional generative adversarial networks." In: arXiv preprint arXiv:1511.06434 (2015).

[73] Aaron Van Den Oord, Nal Kalchbrenner, and Koray Kavukcuoglu. "Pixel recurrent neural networks." In: International conference on machine learning. PMLR. 2016, pp. 1747-1756.

[74] Mayra Macas, Chunming Wu, and Walter Fuertes. "An Attention-Based Deep Generative Model for Anomaly Detection in Industrial Control Systems." In: arXiv preprint arXiv:2405.05277 (2024).

[75] Thomas Schlegl et al. "Unsupervised anomaly detection with generative adversarial networks to guide marker discovery." In: International conference on information processing in medical imaging. Springer. 2017, pp. 146-157.

[76] Jeff Donahue, Philipp Krahenbuhl, and Trevor Darrell. "Adversarial feature learning." In: arXiv preprint arXiv:1605.09782 (2016).

[77] Pascal Vincent et al. "Extracting and composing robust features with de-noising autoencoders." In: Proceedings of the 25th international conference on Machine learning. 2008, pp. 1096-1103.

[78] Augustus Odena. "Semi-supervised learning with generative adversarial networks." In: arXiv preprint arXiv:1606.01583 (2016).

[79] Nicholas Metropolis and Stanislaw Ulam. "The monte carlo method." In: Journal of the American statistical association 44.247 (1949), pp. 335-341.

[80] Radford M Neal. "Probabilistic inference using Markov chain Monte Carlo methods." In: (1993).

[81] Geoffrey E Hinton and Ruslan R Salakhutdinov. "Reducing the dimensionality of data with neural networks." In: science 313.5786 (2006), pp. 504507.

[82] Ruslan Salakhutdinov and Geoff Hinton. "Learning a nonlinear embedding by preserving class neighbourhood structure." In: Artificial intelligence and statistics. PMLR. 2007, pp. 412-419.

[83] Jürgen Schmidhuber. "Learning factorial codes by predictability minimization." In: Neural Computation 4.6 (1992), pp. 863-879.

[84] Were generative adversarial networks introduced by Jürgen Schmidhuber? url: https : / /stats . stackexchange . com / questions / 251460 / were -generative - adversarial - networks - introduced - by - j % C3 % BCrgen -schmidhuber (visited on 12/20/2019).

[85] Mehdi Mirza and Simon Osindero. Conditional Generative Adversarial Nets. 2014. arXiv: 1411.1784 [cs.LG].

[86] Martin Arjovsky and Léon Bottou. "Towards principled methods for training generative adversarial networks." In: arXiv preprint arXiv:1701.04862 (2017).

[87] Tim Salimans et al. "Improved techniques for training gans." In: Advances in neural information processing systems 29 (2016).

[88] Luke Metz et al. "Unrolled generative adversarial networks." In: arXiv preprint arXiv:1611.02163 (2016).

[89] Martin Arjovsky, Soumith Chintala, and Léon Bottou. "Wasserstein Generative Adversarial Networks." In: Proceedings of the 34th International Conference on Machine Learning. Ed. by Doina Precup and Yee Whye Teh. Vol. 70. Proceedings of Machine Learning Research. International Convention Centre, Sydney, Australia: PMLR, Aug. 2017, pp. 214-223. url: http: //proceedings.mlr.press/v70/arjovsky17a.html.

[90] Канторович, Леонид Витальевич and Рубинштейн, Геннадий Шлемович. "Об одном функциональном пространстве и некоторых экстремальных задачах." In: Доклады Академии наук. Vol. 115. 6. Академия наук СССР. 1957, pp. 10581061.

[91] Jinwon An and Sungzoon Cho. "Variational autoencoder based anomaly detection using reconstruction probability." In: Special lecture on IE 2.1 (2015), pp. 1-18.

[92] Alfredo Nazabal et al. "Handling incomplete heterogeneous data using vaes." In: Pattern Recognition 107 (2020), p. 107501.

[93] Chitwan Saharia et al. "Image super-resolution via iterative refinement." In: IEEE transactions on pattern analysis and machine intelligence 45.4 (2022), pp. 4713-4726.

[94] Zhifeng Kong et al. "Diffwave: A versatile diffusion model for audio synthesis." In: arXiv preprint arXiv:2009.09761 (2020).

[95] Prafulla Dhariwal and Alexander Nichol. "Diffusion models beat gans on image synthesis." In: Advances in neural information processing systems 34 (2021), pp. 8780-8794.

[96] Tim Salimans and Jonathan Ho. "Progressive distillation for fast sampling of diffusion models." In: arXiv preprint arXiv:2202.00512 (2022).

[97] Yang Song et al. "Consistency models." In: arXiv preprint arXiv:2303.01469 (2023).

[98] Ziehen Miao, Zhengyuan Yang, Kevin Lin, et al. "Tuning Timestep-Distilled Diffusion Model Using Pairwise Sample Optimization." In: Open-Review preprint (2025). url: https : / / openreview . net / pdf / a461315e8a8df81d012239e10edcdf01d19436b5.pdf.

[99] Robin Rombach et al. "High-resolution image synthesis with latent diffusion models." In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2022, pp. 10684-10695.

[100] Lingxiao Liu et al. "Rectified Flow: Injecting Transforms into Generative Models." In: arXiv preprint arXiv:2302.05858 (2023).

[101] Weihao Zheng et al. "Gaussian Mixture Diffusion Models." In: arXiv preprint arXiv:2401.02840 (2024).

[102] Siddhant Jain, Joseph Geraci, and Harry E Ruda. "Comparing Classical and Quantum Generative Learning Models for High-Fidelity Image Synthesis." In: Technologies 11.6 (2023), p. 183.

[103] M Clemencic et al. "The LHCb simulation application, Gauss: design, evolution and experience." In: Journal of Physics: Conference Series. Vol. 331. 3. IOP Publishing. 2011, p. 032023.

[104] Sea Agostinelli et al. "GEANT4—a simulation toolkit." In: Nuclear instruments and methods in physics research section A: Accelerators, Spectrometers, Detectors and Associated Equipment 506.3 (2003), pp. 250-303.

[105] John Allison et al. "Recent developments in Geant4." In: Nuclear instruments and methods in physics research section A: Accelerators, Spectrometers, Detectors and Associated Equipment 835 (2016), pp. 186-225.

[106] Michela Paganini, Luke de Oliveira, and Benjamin Nachman. "CaloGAN: Simulating 3D high energy particle showers in multilayer electromagnetic calorimeters with generative adversarial networks." In: Physical Review D 97.1 (2018), p. 014021.

[107] B Denby. "Neural networks and cellular automata in experimental high energy physics." In: Computer Physics Communications 49.3 (1988), pp. 429448.

[108] Byron P Roe et al. "Boosted decision trees as an alternative to artificial neural networks for particle identification." In: Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 543.2-3 (2005), pp. 577-584.

[109] Pierre Baldi, Peter Sadowski, and Daniel Whiteson. "Searching for exotic particles in high-energy physics with deep learning." In: Nature communications 5 (2014), p. 4308.

[110] Adam Aurisano et al. "A convolutional neural network neutrino event classifier." In: Journal of Instrumentation 11.09 (2016), P09001.

[111] Steven Farrell et al. "Novel deep learning methods for track reconstruction." In: arXiv preprint arXiv:1810.06111 (2018).

[112] Gregor Kasieczka et al. "The machine learning landscape of top taggers." In: SciPost Physics 7.1 (2019), p. 014.

[113] Theo Heimel et al. "QCD or What?" In: SciPost Physics 6.3 (2019), p. 030.

[114] Luke de Oliveira, Michela Paganini, and Benjamin Nachman. "Learning Particle Physics by Example: Location-Aware Generative Adversarial Networks for Physics Synthesis." In: Computing and Software for Big Science 1.1 (Sept. 2017), p. 4. issn: 2510-2044. doi: 10 . 1007 / s41781 - 017 - 0004 - 6. url: https://doi.org/10.1007/s41781-017-0004-6.

[115] Benjamin Nachman and David Shih. "Anomaly detection with density estimation." In: Physical Review D 101.7 (2020), p. 075042.

[116] Olmo Cerri et al. "Variational autoencoders for new physics mining at the large hadron collider." In: Journal of High Energy Physics 2019.5 (2019), pp. 1-29.

[117] Anja Butter et al. "Machine learning and LHC event generation." In: SciPost physics 14.4 (2023), p. 079.

[118] JM Campbell et al. "Event generators for high-energy physics experiments." In: SciPost Physics 16.5 (2024), p. 130.

[119] S. Agostinelli et al. "Geant4—a simulation toolkit." In: Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 506.3 (2003), pp. 250-303. issn: 0168-9002. doi: https : //doi . org/10 . 1016/S0168-9002(03) 01368-8. url: https : //www.sciencedirect.com/science/article/pii/S0168900203013688.

[120] Alfredo Ferrari et al. "FLUKA: A multi-particle transport code (Program version 2005)." In: (Oct. 2005). doi: 10.2172/877507.

[121] J De Favereau et al. "DELPHES 3: a modular framework for fast simulation of a generic collider experiment." In: Journal of High Energy Physics 2014.2 (2014), pp. 1-26.

[122] Francesco Vaselli et al. FlashSim prototype: an end-to-end fast simulation using Normalizing Flow. Tech. rep. 2023.

[123] Emanuel Parzen. "On estimation of a probability density function and mode." In: The annals of mathematical statistics 33.3 (1962), pp. 1065-1076.

[124] Wolfgang Lukas. "Fast simulation for ATLAS: Atlfast-II and ISF." In: Journal of Physics: Conference Series. Vol. 396. 2. IOP Publishing. 2012, p. 022031.

[125] Douglas Orbaker et al. "Fast simulation of the CMS detector." In: Journal of Physics: Conference Series. Vol. 219. 3. IOP Publishing. 2010, p. 032053.

[126] D Autiero et al. "Parameterization of e and y initiated showers in the NOMAD lead-glass calorimeter." In: Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 425.1-2 (1999), pp. 188-209.

[127] E Barberio et al. "Fast simulation of electromagnetic showers in the ATLAS calorimeter: Frozen showers." In: Journal of Physics: Conference Series. Vol. 160. 1. IOP Publishing. 2009, p. 012082.

[128] Luke de Oliveira, Michela Paganini, and Benjamin Nachman. "Learning particle physics by example: location-aware generative adversarial networks for physics synthesis." In: Computing and Software for Big Science 1.1 (2017), p. 4.

[129] Dalila Salamani et al. "Deep generative models for fast shower simulation in ATLAS." In: 2018 IEEE 14th International Conference on e-Science (e-Science). IEEE. 2018, pp. 348-348.

[130] Viktoria Chekalina et al. "Generative models for fast calorimeter simulation: the LHCb case." In: EPJ Web of Conferences. Vol. 214. EDP Sciences. 2019, p. 02034.

[131] Martin Erdmann, Jonas Glombitza, and Thorben Quast. "Precise simulation of electromagnetic calorimeter showers using a Wasserstein Generative Adversarial Network." In: Computing and Software for Big Science 3.1 (2019), p. 4.

[132] Baran Hashemi. "Deep Generative Models for Ultra-High Granularity Particle Physics Detector Simulation: A Voyage From Emulation to Extrapolation." In: arXiv preprint arXiv:2403.13825 (2024).

[133] Baran Hashemi et al. "Ultra-high-granularity detector simulation with intraevent aware generative adversarial network and self-supervised relational reasoning." In: Nature Communications 15.1 (2024), p. 4916.

[134] Gul rukh Khattak, Sofia Vallecorsa, and Federico Carminati. "Three dimensional energy parametrized generative adversarial networks for electromagnetic shower simulation." In: 2018 25th IEEE International Conference on Image Processing (ICIP). IEEE. 2018, pp. 3913-3917.

[135] Gul Rukh Khattak et al. "Particle detector simulation using generative adversarial networks with domain related constraints." In: 2019 18th IEEE International Conference On Machine Learning And Applications (ICMLA). IEEE. 2019, pp. 28-33.

[136] Dawit Belayneh et al. "Calorimetry with deep learning: particle simulation and reconstruction for collider physics." In: The European Physical Journal C 80.7 (2020), pp. 1-31.

[137] Riccardo Di Sipio et al. "DijetGAN: a generative-adversarial network approach for the simulation of QCD dijet events at the LHC." In: Journal of high energy physics 2019.8 (2019).

[138] S. Diefenbacher et al. "DCTRGAN: improving the precision of generative models with reweighting." In: Journal of Instrumentation 15.11 (Nov. 2020), P11004-P11004. issn: 1748-0221. doi: 10.1088/1748-0221/15/11/p11004. url: http://dx.doi.org/10.1088/1748-0221/15/11/P11004.

[139] Raghav Kansal et al. Graph Generative Adversarial Networks for Sparse Data Generation in High Energy Physics. 2021. arXiv: 2012 . 00173 [physics.data-an]. url: https://arxiv.org/abs/2012.00173.

[140] Sergey Shirobokov et al. "Black-Box Optimization with Local Generative Surrogates." In: Advances in Neural Information Processing Systems. Ed. by H. Larochelle et al. Vol. 33. Curran Associates, Inc., 2020, pp. 14650-14662. url: https : / /proceedings . neurips . cc/paper _ files/paper / 2020 / fiWa878dbebc902328b41dbf02aa87abb58-Paper.pdf.

[141] Claudius Krause and David Shih. "CaloFlow: fast and accurate generation of calorimeter showers with normalizing flows." In: arXiv preprint arXiv:2106.05285 (2021).

[142] Robert D Cousins. "On goodness-of-fit tests." In: Lecture at the University of California (2016).

[143] Raghav Kansal et al. "Evaluating generative models in high energy physics." In: Physical Review D 107.7 (2023), p. 076017.

[144] Alfred Müller. "Integral probability metrics and their generating classes of functions." In: Advances in applied probability 29.2 (1997), pp. 429-443.

[145] Bharath K. Sriperumbudur et al. On integral probability metrics, 0-divergences and binary classification. 2009. arXiv: 0901.2698 [cs.IT]. url: https://arxiv.org/abs/0901.2698.

[146] Bharath K. Sriperumbudur et al. "On the empirical estimation of integral probability metrics." In: Electronic Journal of Statistics 6 (2012), pp. 15501599. url: https://api.semanticscholar.org/CorpusID:14221171.

[147] Aaditya Ramdas, Nicolás García Trillos, and Marco Cuturi. "On Wasserstein Two-Sample Testing and Related Families of Nonparametric Tests." In: Entropy 19.2 (2017). issn: 1099-4300. doi: 10 . 3390/e19020047. url: https://www.mdpi.com/1099-4300/19Z2/47.

[148] Martin Heusel et al. "Gans trained by a two time-scale update rule converge to a local nash equilibrium." In: Advances in neural information processing systems 30 (2017).

[149] Christian Szegedy et al. "Rethinking the inception architecture for computer vision." In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2016, pp. 2818-2826.

[150] Min Jin Chong and David Forsyth. "Effectively unbiased fid and inception score and where to find them." In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020, pp. 6070-6079.

[151] Mikolaj Binkowski et al. "Demystifying mmd gans." In: arXiv preprint arXiv:1801.01401 (2018).

[152] Hosein Hashemi et al. "Ultra-high-resolution detector simulation with intraevent aware GAN and self-supervised relational reasoning." In: arXiv preprint arXiv:2303.08046 (2023).

[153] Riccardo Guidotti et al. "A survey of methods for explaining black box models." In: ACM computing surveys (CSUR) 51.5 (2018), pp. 1-42.

[154] David Lopez-Paz and Maxime Oquab. "Revisiting classifier two-sample tests." In: arXiv preprint arXiv:1610.06545 (2016).

[155] M Faucci Giannelli et al. "Fast calorimeter simulation challenge 2022." In: Online access (16.01. 2024): https://calochallenge. github. io/homepage (2022).

[156] Tuomas Kynkaanniemi et al. "Improved precision and recall metric for assessing generative models." In: Advances in neural information processing systems 32 (2019).

[157] Mehdi SM Sajjadi et al. "Assessing generative models via precision and recall." In: Advances in neural information processing systems 31 (2018).

[158] Raghav Kansal et al. "Particle cloud generation with message passing generative adversarial networks." In: Advances in Neural Information Processing Systems 34 (2021), pp. 23858-23871.

[159] Raffaele Tito d'Agnolo et al. "Learning multivariate new physics." In: The European Physical Journal C 81.1 (2021), pp. 1-21.

[160] Simone Marzani, Gregory Soyez, and Michael Spannowsky. Looking Inside Jets: An Introduction to Jet Substructure and Boosted-object Phenomenology. Springer International Publishing, 2019. isbn: 9783030157098. doi: 10. 1007/978-3-030-15709-8. url: http://dx.doi.org/10.1007/978-3-030-15709-8.

[161] Andrew J Larkoski, Ian Moult, and Benjamin Nachman. "Jet substructure at the Large Hadron Collider: a review of recent advances in theory and machine learning." In: Physics Reports 841 (2020), pp. 1-63.

[162] Ali Borji. "Pros and cons of GAN evaluation measures." In: Computer vision and image understanding 179 (2019), pp. 41-65.

[163] Ali Borji. "Pros and cons of GAN evaluation measures: New developments." In: Computer Vision and Image Understanding 215 (2022), p. 103329.

[164] M. Bernardini and K. Foraz. "Long Shutdown 2 @ LHC." In: CERN Yellow Reports 2.00 (2016), p. 290. url: https://e-publishing.cern.ch/index. php/CYR/article/view/159.

[165] I Bediaga et al. Framework TDR for the LHCb Upgrade: Technical Design Report. Tech. rep. CERN-LHCC-2012-007. LHCb-TDR-12. Apr. 2012. url: https://cds.cern.ch/record/1443882.

[166] Georges Aad et al. "Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC." In: Physics Letters B 716.1 (2012), pp. 1-29.

[167] TL Collaboration et al. "The LHCb detector at the LHC." In: Journal of instrumentation 3.08 (2008), S08005-S08005.

[168] J Schukraft. "Heavy-ion physics with the ALICE experiment at the CERN Large Hadron Collider." In: Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences 370.1961 (2012), pp. 917-932.

[169] G Antchev et al. "The TOTEM detector at LHC." In: Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment 617.1-3 (2010), pp. 62-66.

[170] Esma Mobs. "The CERN accelerator complex - 2019. Complexe des accélérateurs du CERN - 2019." In: (June 2019). General Photo. url: https://cds. cern.ch/record/2684277.

[171] Oliver Sim Brüning et al. "LHC design report, CERN yellow reports: monographs." In: CERN, Geneva (2004).

[172] Lyndon Evans and Philip Bryant. "LHC machine." In: Journal of instrumentation 3.08 (2008), S08001.

[173] Lucio Rossi and Luca Bottura. "Superconducting magnets for particle accelerators." In: Reviews of accelerator science and technology 5 (2012), pp. 5189.

[174] R Aaij et al. "Performance of the LHCb Vertex Locator." In: Journal of Instrumentation 9.09 (Sept. 2014), P09007-P09007. doi: 10.1088/1748-0221/9/09/p09007. url: https : //doi . org/10 . 1088°/.2F1748-02210/02F9°/„ 2F09%2Fp09007.

[175] Roel Aaij et al. "Performance of the LHCb vertex locator." In: Journal of Instrumentation 9.09 (2014), P09007.

[176] LHCb Collaboration. LHCb Silicon Tracker - Material for Publications. url: https://lhcb.physik.uzh.ch/ST/public/material/.

[177] Lucio Anderlini. "Measurement of the B+ meson lifetime using B+ ^ Jdecays with the LHCb detector at CERN." Presented 03 Feb 2015. Jan. 2014. url: http://cds.cern.ch/record/1980460.

178] CMS Collaboration et al. "The CMS ECAL performance with examples." In: JINST 9 (2014), p. C02008.

179] Hermann Kolanoski and Norbert Wermes. Particle Detectors: Fundamentals and Applications. Oxford University Press, USA, 2020.

180] V Daniel Elvira. "Impact of detector simulation in particle physics collider experiments." In: Physics Reports 695 (2017), pp. 1-54.

181] Matthew Feickert and Benjamin Nachman. "A living review of machine learning for particle physics." In: arXiv preprint arXiv:2102.02770 (2021).

182] Eduardo Picatoste Olloqui, LHCb Collaboration, et al. "Lhcb preshower (ps) and scintillating pad detector (spd): Commissioning, calibration, and monitoring." In: Journal of Physics: Conference Series. Vol. 160. 1. IOP Publishing. 2009, p. 012046.

183] Fedor Sergeev et al. "Fast simulation of the LHCb electromagnetic calorimeter response using VAEs and GANs." In: Journal of Physics: Conference Series 1740 (Jan.2021), p. 012028. doI: 10.1088/1742-6596/1740/1/012028.

184] Javier Béjar Alonso. "K-means vs Mini Batch K-means: a comparison." In: (2013).

185] Han Zhang et al. "Self-attention generative adversarial networks." In: International conference on machine learning. PMLR. 2019, pp. 7354-7363.

186] Xiaolong Wang et al. "Non-local neural networks." In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2018, pp. 77947803.

187] Jae Hyun Lim and Jong Chul Ye. "Geometric gan." In: arXiv preprint arXiv:1705.02894 (2017).

188] Bernhard Schölkopf and Alexander J Smola. Learning with kernels: support vector machines, regularization, optimization, and beyond. MIT press, 2002.

189] Takeru Miyato et al. "Spectral normalization for generative adversarial networks." In: arXiv preprint arXiv:1802.05957 (2018).

190] Gene H Golub and Henk A Van der Vorst. "Eigenvalue computation in the 20th century." In: Journal of Computational and Applied Mathematics 123.12 (2000), pp. 35-65.

191] Yuichi Yoshida and Takeru Miyato. "Spectral norm regularization for improving the generalizability of deep learning." In: arXiv preprint arXiv:1705.10941 (2017).

192] John Allison et al. "Geant4 developments and applications." In: IEEE Transactions on nuclear science 53.1 (2006), pp. 270-278.

193] Yiheng Xie et al. "Neural fields in visual computing and beyond." In: Computer Graphics Forum. Vol. 41. 2. Wiley Online Library. 2022, pp. 641-676.

194] Âke Björck and Clazett Bowie. "An iterative algorithm for computing the best estimate of an orthogonal matrix." In: SIAM Journal on Numerical Analysis 8.2 (1971), pp. 358-364.

[195] Cem Anil, James Lucas, and Roger Grosse. "Sorting out Lipschitz function approximation." In: International Conference on Machine Learning. PMLR. 2019, pp. 291-301.

[196] Hsueh-Ti Derek Liu et al. "Learning smooth neural functions via lipschitz regularization." In: ACM SIGGRAPH 2022 Conference Proceedings. 2022, pp. 1-13.

[197] Qiyang Li et al. "Preventing gradient attenuation in lipschitz constrained convolutional networks." In: Advances in neural information processing systems 32 (2019).

[198] Henry Gouk et al. "Regularisation of neural networks by enforcing lipschitz continuity." In: Machine Learning 110 (2021), pp. 393-416.

[199] Mihaela Rosca et al. "A case for new neural network smoothness constraints." In: (2020).

[200] David Terjék. "Adversarial lipschitz regularization." In: arXiv preprint arXiv:1907.05681 (2019).

[201] Andrei N Tikhonov. "Solution of incorrectly formulated problems and the regularization method." In: Sov Dok 4 (1963), pp. 1035-1038.

[202] Robert Tibshirani. "Regression shrinkage and selection via the lasso." In: Journal of the Royal Statistical Society Series B: Statistical Methodology 58.1 (1996), pp. 267-288.

[203] Adrien B Taylor, Julien M Hendrickx, and François Glineur. "Exact worst-case convergence rates of the proximal gradient method for composite convex minimization." In: Journal of Optimization Theory and Applications 178 (2018), pp. 455-476.