Artificial intelligence for high energy physics

목차 (Table of Contents)

• CONTENTS
• Chapter 1. Introduction / Paolo Calafiura, David Rousseau ; Kazuhiro Terao = 1
• 1. Artificial Intelligence at the Frontiers of High-Energy Physics = 1
• 2. Why This Book, and Who Should Read It? = 2
• 3. Pre-requisites = 3
• 4. How to Use This Book = 3
• References = 4
• Part I. Discriminative Models for Signal/Background Boosting = 7
• Chapter 2. Boosted Decision Trees / Yann Coadou = 9
• 1. Introduction = 9
• 2. Specificity of High-Energy Physics = 10
• 3. Decision Trees = 20
• 4. Boosted Decision Trees = 34
• 5. Other Averaging Techniques = 53
• 6. Software = 54
• 7. Conclusion = 55
• References = 55
• Chapter 3. Deep Learning from Four Vectors / Pierre Baldi, Peter Sadowski ; Daniel Whiteson = 59
• 1. Introduction : Pre-Deep Learning State-of-the-Art = 59
• 2. Application of Deep Learning to Four Vectors = 61
• 3. Parameterized Networks = 66
• 4. Handling Sets of Four Vectors = 72
• 5. Physics-aware Networks = 75
• 6. Conclusions = 78
• References = 79
• Chapter 4. Anomaly Detection for Physics Analysis and Less Than Supervised Learning / Benjamin Nachman = 85
• 1. Introduction = 85
• 2. Model Dependence in HEP Data Analysis = 86
• 3. Signal Independent, Background Model Dependent = 91
• 4. Supervised Approaches = 92
• 5. Unsupervised Approaches = 93
• 6. Weak Supervision and Topic Modeling = 95
• 7. Hybrid Approaches = 100
• 8. Results with Collider Data = 102
• 9. Conclusions and Outlook = 104
• References = 106
• Part II. Data Quality Monitoring = 113
• Chapter 5. Data Quality Monitoring Anomaly Detection / Adrian Alan Pol, Gianluca Cerminara, Cecile Germain ; Maurizio Pierini = 115
• 1. Introduction = 116
• 2. Data Quality Monitoring for the LHC Experiments = 117
• 3. Machine Learning Anomaly Detection for HEP DQM = 124
• 4. Detector Components Anomaly Detection with Convolutional Neural Networks and Autoencoders = 127
• 5. Data Certification Novelty Detection with Deep Autoencoders = 134
• 6. Trigger Rate Anomaly Detection with Conditional Variational Autoencoders = 136
• 7. LHC Monitoring with LSTMs = 143
• 8. Conclusion = 146
• References = 146
• Part III. Generative Models = 151
• Chapter 6. Generative Models for Fast Simulation / Michela Paganini, Luke de Oliveira,
• Benjamin Nachman, Denis Derkach, Fedor Ratnikov, Andrey Ustyuzhanin ; Aishik Ghosh = 153
• 1. Generative Models for the Simulation of Particle Showering in Calorimeters = 153
• 2. Conclusions and Outlook = 179
• References = 180
• Chapter 7. Generative Networks for LHC Events / Anja Butter ; Tilman Plehn = 191
• 1. Introduction = 191
• 2. Generative Networks = 196
• 3. Neural Networks in Event Generators = 204
• 4. GANs and VAEs as Event Generators = 215
• 5. Inverting the Simulation Chain = 227
• 6. Outlook = 234
• References = 236
• Part IV. Machine Learning Platforms = 241
• Chapter 8. Distributed Training and Optimization of Neural Networks / Jean-Roch Vlimant ; Junqi Yin = 243
• 1. Introduction = 243
• 2. Neural Network Optimization Formalism = 245
• 3. Parameter Distribution = 248
• 4. Data Distributed Training = 250
• 5. Model Parallelism = 253
• 6. Hyperparameter Optimization = 255
• 7. Summary and Discussion = 258
• References = 261
• Chapter 9. Machine Learning for Triggering and Data Acquisition / Philip Harris ; Nhan Tran = 265
• 1. Introduction = 265
• 2. Fast ML for Real-Time Readout and Near-Detector Triggering = 271
• 3. Fast ML for Data Processing = 280
• 4. Concluding Remarks = 300
• References = 301
• Part V. Detector Data Reconstruction = 311
• Chapter 10. End-to-End Analyses Using Image Classification / Adam Aurisano ; Leigh H. Whitehead = 313
• 1. Introduction = 313
• 2. Traditional Workflow = 314
• 3. Deep Learning Approaches = 315
• 4. Convolutional Neural Networks in Lattice-Structured Experiments = 324
• 5. Convolutional Neural Networks in Heterogeneous Collider Detectors = 335
• 6. End-to-End Analysis of Time Series Using One-Dimensional CNNs = 338
• 7. Graph Neural Networks for Large Three-Dimensional Detectors = 340
• 8. Opening the Black-Box = 341
• 9. Conclusions = 349
• References = 350
• Chapter 11. Clustering / Kazuhiro Terao = 355
• 1. Introduction = 355
• 2. Fixed Number of Partitions = 359
• 3. Convolutional Neural Networks for Pixel Clustering = 368
• 4. Clustering Particles Using Graph Neural Networks = 375
• 5. Summary = 381
• References = 382
• Chapter 12. Graph Neural Networks for Particle Tracking and Reconstruction / Javier Duarte ; Jean-Roch Vlimant = 387
• 1. Introduction = 387
• 2. Point Cloud and Graph Data = 391
• 3. Graph Neural Networks = 396
• 4. GNN Design Considerations = 402
• 5. Applications to Particle Physics Tasks = 406
• 6. Summary = 423
• References = 425
• Part VI. Jet Classification and Particle Identification from Low Level = 437
• Chapter 13. Image-Based Jet Analysis / Michael Kagan = 439
• 1. Introduction = 439
• 2. Jets and Jet Physics Challenges = 442
• 3. Jet Images and Preprocessing = 445
• 4. Computer Vision and Convolutional Neural Networks = 449
• 5. Jet Tagging = 454
• 6. Understanding Jet Image-Based Tagging = 480
• 7. Other Applications of Jet Images = 485
• 8. Conclusion = 491
• References = 492
• Chapter 14. Particle Identification in Neutrino Detectors / Ralitsa Sharankova ; Taritree Wongjirad = 497
• 1. Introduction = 497
• 2. Behavior of Particles in Matter = 498
• 3. Neutrino Interactions with Matter = 501
• 4. Scintillator Detectors = 502
• 5. Cherenkov Ring Imaging Detectors = 513
• 6. Tracking Detectors = 519
• 7. Concluding Thoughts = 536
• References = 538
• Chapter 15. Sequence-Based Learning / Rafael Teixeira de Lima = 541
• 1. Introduction = 541
• 2. Applications of RNNs to Jet Physics = 546
• 3. Alternatives to RNNs = 565
• 4. Conclusion = 572
• References = 573
• Part VII. Physics Inference = 577
• Chapter 16. Simulation-Based Inference Methods for Particle Physics / Johann Brehmer ; Kyle Cranmer = 579
• 1. Particle Physics Measurements as a Simulation-Based Inference Problem = 579
• 2. Inference with Surrogates = 586
• 3. Inference with Sufficient Summary Statistics = 594
• 4. Diagnostics, Calibration, and Systematic Uncertainties = 597
• 5. Probabilistic Programming = 599
• 6. Software and Computing = 602
• 7. Summary = 604
• References = 605
• Chapter 17. Dealing with Nuisance Parameters / T. Dorigo ; P. de Castro Manzano = 613
• 1. Introduction = 613
• 2. Nuisance-Parameterized Models = 626
• 3. Feature Decorrelation, Penalized Methods, and Adversary Losses = 629
• 4. Semi-supervised Approaches = 638
• 5. Inference-Aware Approaches = 642
• 6. Outlook = 657
• References = 657
• Chapter 18. Bayesian Neural Networks / Tom Charnock, Laurence Perreault-Levasseur ; François Lanusse = 663
• 1. Introduction = 664
• 2. Bayesian Neural Networks = 667
• 3. Practical Implementations = 680
• 4. Concluding Remarks and Outlook = 710
• References = 711
• Chapter 19. Parton Distribution Functions / Stefano Forte ; Stefano Carrazza = 715
• 1. Introduction = 715
• 2. The State of the Art = 724
• 3. The Future of PDFs in a Deep Learning Framework = 737
• References = 759
• Part VIII. Scientific Competitions and Open Datasets = 763
• Chapter 20. Machine Learning Scientific Competitions and Datasets / David Rousseau ; Andrey Ustyuzhanin = 765
• 1. Introduction = 765
• 2. HiggsML = 767
• 3. Flavor of Physics = 772
• 4. TrackML = 782
• 5. LHC Olympics = 793
• 6. Competitions Platforms = 797
• 7. Open Datasets and Responsitories = 805
• 8. Guidelines for New Competition Organizers = 806
• 9. Conclusion = 808
• References = 809
• Index = 813