Buch, Englisch, 448 Seiten, Format (B × H): 152 mm x 229 mm, Gewicht: 769 g
Buch, Englisch, 448 Seiten, Format (B × H): 152 mm x 229 mm, Gewicht: 769 g
ISBN: 978-1-394-21538-6
Verlag: John Wiley & Sons
Authoritative and highly comprehensive resource on the latest research and strategies to develop cyber resilience in any network system
Autonomous Cyber Resilience presents key research contributions in the fields of cyber resilience, resilient machine learning, and game theory for network security. It introduces basic concepts on resilience assessment framework, human robot teaming, zero-trust cyber resilience, the Stackelberg network game, and adversarial machine learning. The book describes a comprehensive suite of solutions for a broad range of technical challenges in autonomous cyber resilience, examines network robustness, planning, learning, and self-adaptation in a dynamic and uncertain environment and provides a joint analysis of cyber resilience and machine learning resilience.
The book gathers experts in this emerging area of research to share their latest contributions in federated learning, resilient deep neural networks, topological data analysis, and effective deployment of honeypots, with valuable insights on applying these new methods to address cyber autonomy, network intrusion detection, and NextG communication systems. Additional chapters summarize ongoing research topics in cyber security and point to open issues and future research challenges and opportunities for academia and industry.
Autonomous Cyber Resilience includes information on: - Hypergraphs as a tool to move beyond basic pairwise relations and interactions to accurately model higher order interactions between groups of agents
- Settings where multiple, distributed, and collaborative bots involved in an attack can make the impact of vulnerabilities more severe
- The Resilience Index, the percentage of Monte Carlo simulations where mission essential functions perform below the acceptable threshold
- Eigenvector centrality, a metric that takes into account not just the centrality (degree) of a node but also its power
Providing an extensive set of techniques to meet a diverse array of obstacles in the field, Autonomous Cyber Resilience is essential reading for researchers, students, and experts in the fields of computer science and engineering, along with industry and military professionals involved in projects related to cybersecurity.
Autoren/Hrsg.
Fachgebiete
Weitere Infos & Material
Preface xv
Editor Biographies xvii
1 Introduction 1
Alexander Kott
1.1 Cyber Resilience and Cybersecurity 1
1.2 Autonomy and Cyber Resilience 3
1.3 Autonomous Actions 4
1.4 Approaches to Implementing Autonomous Cyber Resilience 5
1.5 Benefits and Risks of Autonomous Cyber Resilience 8
1.6 The Preview of the Book 9
1.6.1 Part I: Foundations of Cyber Resilience 9
1.6.2 Part II: Resilient Machine Learning 10
1.6.3 Part III: Game Theory for Network Resilience 10
References 11
Part 1 Cyber Resilience 13
2 Game-theoretic Foundations for Cyber Resilience Against Deceptive Information Attacks in Intelligent Transportation Systems 15
Ya-Ting Yang and Quanyan Zhu
2.1 Introduction 15
2.1.1 Multi-domain Threats in ITS 17
2.1.2 Security Risk Assessment 18
2.1.3 Chapter Organization 19
2.2 Deceptive Information Attacks 19
2.2.1 Intra-vehicle Domain 20
2.2.2 Inter-vehicle Domain Attacks 22
2.2.3 Transportation System Domain 23
2.2.4 Human Aspects 27
2.3 Cross-layer Resilience 29
2.3.1 Holistic Framework for Resilience 30
2.3.2 Theoretical Foundations and Design Frameworks 31
2.3.3 Benefits of Game-theoretic and Learning-based Design Principles for Cyber Resilience 34
2.4 Case Study 35
2.4.1 Misinformation Attacks on Recommendation Systems 35
2.5 Conclusion and Discussion 43
2.5.1 Conclusion 43
2.5.2 Discussion 43
References 44
3 CYBER-MIRA: Cyber Mission Impact Resilience Assessment Framework for Tactical Mission Systems 53
Ashrith Reddy Thukkaraju, Han Jun Yoon, Shou Matsumoto, Jair Feldens Ferrari, Donghwan Lee, Myung Kil Ahn, Paulo Costa, and Jin-Hee Cho
3.1 Introduction 53
3.1.1 Motivation and Challenges 54
3.1.2 Research Goal 55
3.1.3 Key Contributions 55
3.1.4 Structure of This Chapter 56
3.2 Related Work 56
3.2.1 Methodologies of CMIA 56
3.2.2 MIA Frameworks 57
3.2.3 Cyber Resilience Assessment 58
3.3 System Model 59
3.3.1 Network Model 60
3.3.2 Attack Model 60
3.3.3 Defense Model 62
3.4 CYBER-MIRA Framework 64
3.4.1 Architecture of CYBER-MIRA 65
3.4.2 Hypergame Expected Utility 74
3.4.3 Resilience Assessment as a Measure of Performance 82
3.5 Limitations 86
3.6 Conclusion and Future Work 86
3.6.1 Summary of the Key Contributions 86
3.6.2 Future Work 87
References 87
4 Modeling Autonomous Network Resilience in Adversarial Environments Using Machine Learning and Topological Data Analysis 91
Nandi O. Leslie
4.1 Introduction 91
4.2 TDA Concepts 93
4.2.1 Simplices, Simplicial Complexes, and Filtration 93
4.2.2 Persistent Homology 95
4.3 Network Resilience Modeling 97
4.4 Conclusion 101
References 101
5 Game-theoretic Frameworks for Zero-trust Authentication in Autonomous Cyber Resilience 105
Yunfei Ge and Quanyan Zhu
5.1 Introduction 105
5.2 From Traditional Security to Zero Trust 108
5.3 Trust Evaluation Design 110
5.3.1 Target 110
5.3.2 Metric 110
5.3.3 Collection and Evaluation 111
5.3.4 Purpose 115
5.3.5 Management 115
5.4 Policy Engine Design 116
5.4.1 Authentication Layer: Continuous Authentication 117
5.4.2 Authorization Layer: Least Privilege Access 117
5.4.3 Network Layer: Microsegmentation 117
5.5 Zero Trust for Cyber Resilience 118
5.5.1 How Zero Trust Contributes to Cyber Resilience 118
5.5.2 A Running Example 120
5.6 Strategic Zero-trust Implementation 123
5.6.1 A Game-theoretic Approach 123
5.6.2 Case Studies 124
5.7 Conclusion 132
References 134
6 Cyber Insurance for Cyber Resilience 139
Shutian Liu and Quanyan Zhu
6.1 Introduction 139
6.2 Attack Models and Insured Targets 144
6.2.1 Human-layer Attacks 144
6.2.2 Cyber-layer Attacks 146
6.2.3 Physical-layer Attacks 148
6.3 Defense Mechanisms and Residual Risks 149
6.3.1 Modeling of Defense Mechanisms 149
6.3.2 Types of Security Investments 150
6.3.3 Residual Risk and Its Connection with Cyber Insurance 153
6.4 Insurer’s Observations and the P-A Model 155
6.4.1 User Behavior Monitoring 155
6.4.2 Principal-agent Problems 156
6.5 Modeling of Risk Preferences 158
6.5.1 Risk Modeling 159
6.5.2 Enhancing Cyber Resilience 161
6.6 Insurance Design with Preference Manipulation 162
6.7 Dynamic Insurances 164
6.8 Regulations on Cyber Insurance 166
6.8.1 Mandatory Cyber Insurance 166
6.8.2 Insurance Policy Elaboration 167
6.8.3 Designing Accountability Mechanisms 168
6.8.4 Insurance Market Monitoring 168
6.9 Conclusion 170
References 170
7 Enhancing Cyber Resiliency: Assessing the Effectiveness of Deploying Honeypots in Different Network Topologies 183
7.1 Introduction 183
7.1.1 Contribution 184
7.2 Task Description 185
7.2.1 Experiment Conditions 185
7.2.2 Experiment Scenario 187
7.2.3 Results 189
7.2.4 Scanning and Exploitation Behavior 189
7.2.5 Operating System and Exploit Preference 192
7.3 A Cognitive Model of Attackers in HackIT 193
7.3.1 IBL Theory 193
7.3.2 IBL Model for Attacker 195
7.3.3 IBL Model Results 197
7.4 Discussion 199
References 200
Part 2 Resilient Machine Learning 203
8 Computational Game Theory for Security 205
Yevgeniy Vorobeychik
8.1 Introduction 205
8.2 Stackelberg Games 206
8.3 Stackelberg Security Games 209
8.4 Security Games on Networks 212
8.5 Stochastic Stackelberg Games and Adversarial Patrolling 215
8.5.1 Stochastic Discounted Stackelberg Games 215
8.5.2 Adversarial Patrolling Games 217
8.5.3 Solving Zero-sum APGs 218
8.5.4 Solving General-sum APGs 220
8.6 Conclusion 221
References 221
9 Privacy and Robustness Trade-offs of Artificial Intelligence Models with Federated Learning 225
Kemal Davaslioglu, Yi Shi, and Yalin E. Sagduyu
9.1 Introduction 225
9.2 Model Inversion Attacks 229
9.2.1 Softmax Regression Model Training 229
9.2.2 ModInv Attacks Against a Single-layer Convolutional Neural Network (CNN) 231
9.2.3 ModInv Attacks Against a Four-layer CNN 231
9.2.4 Comparison of the Three Types of Models Under Attack 232
9.2.5 Privacy Preservation Evaluation of Image Transformations Against ModInv Attacks 233
9.2.6 Accuracy of These Privacy-preserving Image Transformations in ModInv Attacks 241
9.2.7 Demonstration of ModInv Attacks Against CIFAR- 10
Dataset 242
9.3 Membership Inference Attacks 243
9.3.1 Introduction 243
9.3.2 General MI Attack Model 246
9.3.3 Logistic Attack Models 247
9.3.4 Overview of the MI Attack 247
9.3.5 Naïve Attacks 248
9.3.6 The Threat Models 248
9.3.7 Naïve Bayes mi 248
9.3.8 mi in Deep Models 249
9.3.9 Differentially Private Stochastic Gradient Descent for Privacy 252
9.3.10 Design Defense Approaches for MI Attack 253
9.3.11 Membership Privacy in ml 254
9.4 Federated Learning 264
9.4.1 Federated Learning Implementation 265
9.4.2 Aircraft Classification in the xView Dataset Using FL 266
9.4.3 Effect of FL Parameters 268
9.4.4 Demonstration of ModInv Attack Against FL Models 270
9.4.5 Differentially Private Stochastic Gradient Descent 271
9.4.6 Evaluate Effects of Different Parameters of DP-SGD 272
9.4.7 Demonstration of Renyi DP Evaluations 273
9.5 Discussion 275
9.6 Conclusions 276
9.7 Acknowledgment 277
References 277
10 Resilient Deep Neural Network Random Ensemble Against Adversarial Attacks 281
Kirsen Sullivan, Yitao Li, Charles A. Kamhoua, and Bowei xi
10.1 Introduction 281
10.1.1 Related Work 282
10.2 Data and Bootstrapped CNNs 284
10.2.1 Bootstrap Three-layer CNN 285
10.2.2 Bootstrap VGG 16 287
10.2.3 Bootstrap Inception V 3 288
10.3 Bootstrapped Distributions 290
10.3.1 CNN3 Parameter 290
10.3.2 VGG16 Parameters 292
10.3.3 Inception v3 Parameters 293
10.3.4 Normality Test 296
10.3.5 NNs by Varying Initial Random Seeds 297
10.3.6 Regression for NN Parameters 299
10.4 Randomized DNN Ensembles with Gaussian Random Weights 300
10.4.1 Adversarial Examples Generation 300
10.4.2 Randomization 301
10.5 Ensemble Experiment Results 302
10.5.1 CNN3 Randomization Results 302
10.5.2 VGG16 Randomization Results 303
10.5.3 Inception v3 Randomization Results 304
10.6 Conclusion 311
References 312
Part 3 Game Theory for Network Resilience 317
11 Poisoning Attack and Defense Game for Federated Learning in Resilient NextG Networks 319
Yalin E. Sagduyu, Tugba Erpek, and Yi Shi
11.1 Introduction 319
11.2 Federated Learning for Distributed Spectrum Monitoring 323
11.3 Attack and Defense Mechanisms for Resilient Federated Learning 325
11.4 Poisoning Attack–Defense Game for Two Clients 330
11.5 Poisoning Attack–Defense Game for More than Two Clients 337
11.6 Future Research Directions 338
11.7 Conclusion 341
References 341
12 Self-adapting Quantum Network Provisioning Using Game Theory 347
Stefan Rass, Miralem Mehic, Sandra König, Stefan Schauer, and Miroslav Voznak
12.1 Introduction 347
12.2 Basics of Quantum Networks 348
12.3 Game Theory to Orchestrate Cryptography 349
12.3.1 Basics of Perfectly Secure Message Transmission 350
12.3.2 Hierarchical Secret Sharing and Access Structures 352
12.3.3 Using Secret Sharing for Perfectly Secure Multipath Transmission (Defense Strategies) 353
12.3.4 How to Define or Identify Adversary Structures (Attack Strategies)? 354
12.3.5 Quantum Cryptography in Combination with Multipath Transmission 354
12.3.6 Game-theoretic Orchestration of Perfectly Secure Message Transmission 355
12.4 Self-adaption of QKD Devices to Environmental Conditions 361
12.5 Adapting the Level of Service to Traffic Changes 369
12.6 Adapting the Network Topology 372
12.7 Conclusions, Outlook, and QKD in Today’s Networks 376
References 377
13 Conclusion and Future Works 383
Quanyan Zhu
13.1 Overview 383
13.2 Summary 384
13.2.1 Summary and Synthesis of Part 1 on Foundations of Cyber Resilience 385
13.2.2 Summary and Synthesis of Part 2 on Resilient ml 390
13.2.3 Summary of Part 3 on Game Theory for Network Resilience 393
13.3 Future Directions: Charting the Path Toward Autonomous Cyber Resilience 396
13.3.1 Learning Autonomy: From Reactive Defense to Strategic Adaptation 396
13.3.2 Multiscale Resilience: Tailoring Autonomy Across Layers and Contexts 397
13.3.3 Game-theoretic Intelligence: Strategic Reasoning for Resilience 399
13.3.4 Holistic and Integrative Approaches: Building a Converged Resilience Architecture 401
13.3.5 Research and Policy Translation: From Innovation to Implementation 403
13.4 Concluding Remarks 404
Index 407
