[Lecture Notes in Computer Science] Decision and Game Theory for Security Volume 12513 (11th International Conference, GameSec 2020, College Park, MD, USA, October 28–30, 2020, Proceedings) ||

This book constitutes the refereed proceedings of the 11th International Conference on Decision and Game Theory for Security, GameSec 2020,held in College Park, MD, USA, in October 2020. Due to COVID-19 pandemic the conference was held virtually The 21 full papers presented together with 2 short papers were carefully reviewed and selected from 29 submissions. The papers focus on machine learning and security; cyber deception; cyber-physical systems security; security of network systems; theoretic foundations of security games; emerging topics. Preface Organization Contents Machine Learning and Security Distributed Generative Adversarial Networks for Anomaly Detection 1 Introduction 2 Related Work 2.1 Anomaly Detection 2.2 GANs for Anomaly Detection 3 Game-Theoretic Model of Generative Adversarial Networks 3.1 GAN Games 3.2 GAN Training Methods and Challenges 4 A Novel Distributed GAN Framework 4.1 Peer-GAN Game 4.2 Peer-GAN Distributed Training and Convergence 5 Anomaly Detection and Simulation Results 5.1 Peer-GAN Convergence 5.2 Sample Reconstruction 5.3 Anomaly Detection Comparison 6 Conclusion References Learning and Planning in the Feature Deception Problem 1 Introduction 2 The Feature Deception Problem 3 Learning the Adversary's Preferences 4 Computing the Optimal Feature Configuration 5 Experiments 5.1 Learning 5.2 Planning 5.3 Combining Learning and Planning 5.4 Case Study: Credit Bureau Network 6 Related Work 7 Discussion References A Realistic Approach for Network Traffic Obfuscation Using Adversarial Machine Learning 1 Introduction 2 Motivation and Related Work 3 Experimental Setup 3.1 Dataset 3.2 Realistic Features 3.3 Classification Model 4 Adversarial Settings 4.1 Defender Model 4.2 Adversary Model 4.3 Obfuscation Approaches 5 Restricted Traffic Distribution Attack 5.1 Perturbation Constraints 5.2 Distribution Constraints 5.3 Framework 6 Results 7 Conclusions and Future Work References Adversarial Deep Reinforcement Learning Based Adaptive Moving Target Defense 1 Introduction 2 Preliminaries 2.1 Independent Reinforcement Learning 2.2 Deep-Q-Network Learning 3 Model 3.1 Environment and Players 3.2 State 3.3 Actions 3.4 Rewards 3.5 Observations 4 Problem Formulation 4.1 Pure Strategy 4.2 Mixed Strategy 4.3 Solution Concept 5 Framework 5.1 Solution Overview 5.2 Challenges 5.3 Solution Approach 6 Evaluation 6.1 Baseline Heuristic Strategies 6.2 Implementation 6.3 Numerical Results 7 Related Work 7.1 Moving Target Defense 7.2 Reinforcement Learning for Cybersecurity 8 Conclusion References Lie Another Day: Demonstrating Bias in a Multi-round Cyber Deception Game of Questionable Veracity 1 Introduction 2 Related Work 3 Analysis of Optimal Attacker Strategies 3.1 Game Model 3.2 Cost Hypothesis 3.3 Analysis 4 Simulation 4.1 Simulation Model 4.2 Parameters 4.3 Results 4.4 Demonstrating Cognitive Bias 5 Conclusion References Cyber Deception Exploiting Bounded Rationality in Risk-Based Cyber Camouflage Games 1 Introduction 1.1 Related Work 2 Risk-Based Cyber Camouflage Games (RCCG) Model 3 Rational Attackers 3.1 Zero-Sum RCCG 3.2 Unconstrained General-Sum RCCG 3.3 Constrained General-Sum RCCG 4 A Model-Driven Approach with Prospect Theory 4.1 Learning Model Parameters from Data 4.2 Robust Solution with Prospect Theory 5 GEBRA: Exploiting Bounded Rationality Model-Free 6 Numerical Results 7 Summary A RCCG for Rational Attackers B Sensitivity to Learning Error C Computing Strict Competitiveness References Farsighted Risk Mitigation of Lateral Movement Using Dynamic Cognitive Honeypots 1 Introduction 1.1 Related Works 1.2 Notation and Organization of the Paper 2 Chronological Enterprise Network Model 2.1 Time-Expanded Network and Random Service Links 2.2 Attack Model of Lateral Movement over a Long Duration 2.3 Cognitive Honeypot 3 Farsighted Vulnerability Mitigation for Long-Term Security 3.1 Imminent Vulnerability 3.2 k-stage Vulnerability 3.3 Curse of Multiple Attack Paths and Two Sub-Optimal Honeypot Policies 3.4 LTV Analysis Under Two Heuristic Policies 4 Conclusion References Harnessing the Power of Deception in Attack Graph-Based Security Games 1 Introduction 2 Game Model 3 Theoretical Analysis 4 The MILP Approach for Layered DAGs 4.1 Bipartite DAG 4.2 Layered DAG 5 The NAS Approach for General DAGs 6 Experiments 6.1 Bipartite DAGs 6.2 General DAGs 7 Related Work 8 Discussion and Conclusion References Decoy Allocation Games on Graphs with Temporal Logic Objectives 1 Introduction 2 Problem Formulation 2.1 Attack-Defend Games on Graph 2.2 Formulating the Decoy Allocation Problem 3 Main Result 3.1 Deceptive Synthesis: Hypergames on Graphs 3.2 Compositional Synthesis for Decoy Allocation 4 Conclusion References Popular Imperceptibility Measures in Visual Adversarial Attacks are Far from Human Perception 1 Introduction 2 Study Overview 2.1 Human Just Noticeable Difference (JND) 2.2 Popular Imperceptibility Measures d() 3 Human JND Experiments 4 Results 4.1 Qualitative Assessment 4.2 Quantitative Assessment 5 Discussions and Conclusion References Cyber-Physical System Security Secure Discrete-Time Linear-Quadratic Mean-Field Games 1 Introduction 1.1 Agent Model and Objective 1.2 A Motivating Application 1.3 Main Results and Organization 2 Secure LQ-MFG: Model and Objective 2.1 Secure n-agent Linear Quadratic (LQ) Game 2.2 Secure Linear Quadratic Mean-Field Game (SLQ-MFG) 3 State Reconstruction Using Multi-rate Sensor Output Sampling 4 Equilibria of Secure LQ Games 4.1 MFE of the SLQ-MFG 4.2 -MFE of the SLQ-MFG 4.3 (+ )-Nash Equilibrium of the Secure n-Agent LQ Game 4.4 Summary and Discussion 5 Empirical Studies 5.1 Performance Sensitivity w.r.t. Sampling Rate 5.2 Performance Sensitivity w.r.t. Model Parameters and Private Keys 6 Conclusion 7 Appendix 7.1 MFE of the LQ-MFG References Detection of Dynamically Changing Leaders in Complex Swarms from Observed Dynamic Data 1 Introduction 2 Modeling Complex Swarm Maneuvers 2.1 Extended Boids Model 2.2 Cucker-Smale Model with Leadership 3 Leader Detection 3.1 Granger Causality 3.2 Leader Detection Based on Granger Causality 4 Estimating the Number of Leaders 4.1 Deterministic Annealing 5 Learning the Particle Interaction Laws 6 Experimental Results 6.1 Case of One Leader 6.2 Case of Multiple Leaders 7 Conclusion and Discussion References Moving Target Defense for Robust Monitoring of Electric Grid Transformers in Adversarial Environments 1 Introduction 2 Preliminaries 2.1 The Electric Power Grid as a Graph 2.2 Minimum Discriminating Code Set (MDCS) 2.3 Moving Target Defense (MTD) and Differential Immunity 3 K Differentially Immune MDCS (K-MDCS) 3.1 Finding Max K for K-MDCS 4 Game Theoretic Formulation 5 Experimental Simulation 6 Related Works 7 Conclusion References Security of Network Systems Blocking Adversarial Influence in Social Networks 1 Introduction 2 Problem Formulation 3 Solution Approach 3.1 Computing Attacker's Best Response 3.2 Optimal Influence Blocking: A Constraint Generation Approach 3.3 Approximating Optimal Influence 3.4 Scaling up Through a Pruning Heuristic 4 Extensions 5 Experiments 6 Conclusion References Normalizing Flow Policies for Multi-agent Systems 1 Introduction 2 Background and Related Work 2.1 Flow Models 2.2 Stochastic Games 2.3 Imitation Learning and Agent Modeling 2.4 Multi-agent Reinforcement Learning 3 Normalizing Flow Policy Representation 3.1 Conditional Flow as Policy Representation 3.2 Representation Capability 4 Experiments 4.1 Agent Modeling 4.2 Multi-agent RL 5 Conclusion A Appendix References A Game Theoretic Framework for Software Diversity for Network Security 1 Introduction 2 Related Work 3 System Model 3.1 Attacker Problem 3.2 Defender Problem 3.3 Payoff Functions 3.4 Game Problem 3.5 Game Complexity 3.6 Complexity Reduction 4 Numerical Results 5 Conclusion and Future Work References Partially Observable Stochastic Games for Cyber Deception Against Network Epidemic 1 Introduction 2 Model Description 2.1 Problem Description 2.2 Model 2.3 Model Description 3 Solution Description 3.1 Strategies 3.2 Utility 3.3 Objectives 4 Value Backup Operator 5 Numerical Illustrations 6 Conclusions and Further Work References Combating Online Counterfeits: A Game-Theoretic Analysis of Cyber Supply Chain Ecosystem 1 Introduction 1.1 Related Work 1.2 Organization of the Paper 2 Model of Cyber Supply Chain 2.1 Consumers' Model 2.2 Pricing Mechanisms in Licit and Illicit Market 2.3 Sellers' Utility 3 Problem Formulation and Game Structure 3.1 Stackelberg Game 3.2 Nash Game 4 Analysis of Counterfeiting in Cyber Supply Chain 4.1 Market Share Analysis 4.2 Best Response Functions 4.3 Iterative Algorithm 5 Case Studies and Simulations 5.1 Best Response and Equilibrium Strategy 5.2 Discussion on Parameter Sensitivity 5.3 Anti-Counterfeit Strategies 6 Conclusion A Proof of Theorem 1 B Proof of Theorem 2 References Theoretic Foundations of Security Games On the Characterization of Saddle Point Equilibrium for Security Games with Additive Utility 1 Introduction 2 Problem Formulation: Security Game 3 Structural Properties of the Attacker's Strategy 4 Computation of v* 5 Dual Analysis: Structural Properties of the Defender's Strategy and Algorithms 6 Conclusion References MASAGE: Model-Agnostic Sequential and Adaptive Game Estimation 1 Introduction 2 Related Work 3 Problem Formulation 3.1 Preliminary 3.2 Dynamic Linear Estimation Problem 4 Objective Function Analysis 4.1 Basic Properties 4.2 Perturbation Theory of Parameterized Matrix Game 5 Algorithmic Analysis 5.1 Sequential Observation and Adaptation 5.2 Extended Kalman Filter 6 Case Study 6.1 Experimental Setting and Results 6.2 Discussions 7 Conclusions and Future Research References Using One-Sided Partially Observable Stochastic Games for Solving Zero-Sum Security Games with Sequential Attacks 1 Introduction 2 Technical Background 2.1 One-Sided Partially Observable Stochastic Games (OS-POSG) 2.2 Heuristic Search Value Iteration (HSVI) 2.3 Security Games with Sequential Attacks 3 Using OS-POSGs for Sequential Attacks 3.1 Representing SGSA as OS-POSG 3.2 HSVI-Inspired Algorithm 3.3 Exact Variant of the Algorithm 3.4 Heuristic Variant of the Algorithm 4 Experimental Evaluation 4.1 Experiments Setting 4.2 Comparison with State of the Art 4.3 Algorithm Scalability 4.4 Solution Quality 5 Conclusion References A Data-Driven Distributionally Robust Game Using Wasserstein Distance 1 Introduction 2 Preliminaries 3 Robust Game 4 Data-Driven Empirical Game 5 Data-Driven Distributionally Robust Game 5.1 Motivation 5.2 Equilibrium Concept 5.3 Existence of DRE 5.4 Asymptotic Consistency 5.5 Tractable Formulations 5.6 Mathematical Programming for DRE 6 Numerical Example 7 Conclusions and Future Work 7.1 Conclusions 7.2 Future Work References Security Games over Lexicographic Orders 1 Introduction and Motivation 1.1 Related Work 1.2 Our Contribution 2 Preliminaries 2.1 Notation 2.2 Representability of the Lexicographic Order 3 Finding Lex-Order Optimal Strategies 4 Applications and Examples 4.1 Refining Ambiguous Attack or Defense Strategies 4.2 Example 1: The Pure Algorithm (Numerical Illustration) 4.3 Example 2: Data Download 5 Conclusion A Proof of Proposition 1 B Proof of Lemma 2 B.1 Example B.2 Restricting the Equilibria to the Desired Set References Emerging Topics Game Theory on Attack Graph for Cyber Deception 1 Introduction 2 Related Work 2.1 Attack Graph 2.2 Hypergame 3 Information Model and Game Formulation 3.1 On the Complexity of POSG 3.2 Game Model 4 Game Solving Techniques 4.1 POMDP Embedded Game 4.2 One-sided POSG 5 Results 6 Conclusion References Attacking Machine Learning Models for Social Good 1 Introduction 2 Related Work 3 Modeling Socially Good Adversarial Attacks 3.1 Ethical and Practical Adversarial Attacks for Image Classification 3.2 Ethical and Practical Adversarial Attacks for Classification with Discrete Attributes 4 Experiments 4.1 Methodology and Experimentation for CelebA Dataset 4.2 Methodology and Experimentation for German Credit Dataset 5 Conclusion References A Review of Multi Agent Perimeter Defense Games 1 Introduction 2 Problem Statement 3 Solution Method 3.1 Agent-Level Control Policy 3.2 Team-Level Coordination Policies 4 Extensions and Generalizations 4.1 Assignment-Based Defense Policies 4.2 Cooperative Intruder Strategies 5 Limitations and Future Directions 6 Conclusion References Hardware Security and Trust: A New Battlefield of Information 1 Introduction 2 The Role of Hardware in Cybersecurity 3 Key Problems in Hardware Security and Trust 3.1 Physical Attacks 3.2 Side Channel Analysis 3.3 Intellectual Property Protection 3.4 Hardware Trojan 3.5 Hardware Security Primitives 3.6 Applications in Security and Trust 4 Information Battle Perspective of Hardware Security 5 Open Problems and Conclusions Appendix References Security Games with Insider Threats 1 Introduction 2 Private Information 3 Stealthy Attacks with Insider Information 4 Conclusions and Future Research References Securing Next-Generation Wireless Networks: Challenges and Opportunities (Extended Abstract) 1 Introduction 2 Security Goals and Challenges 2.1 Goals 2.2 Challenges 3 Opportunities References Short Paper A Data Mining Friendly Anonymization Scheme for System Logs using Distance Mapping 1 Motivation 2 Problem and Proposed Solution References Author Index