Decision and Game Theory for Security: 12th International Conference, GameSec 2021, Virtual Event, October 25–27, 2021, Proceedings (Lecture Notes in Computer Science)

Description

This book constitutes the refereed proceedings of the 12th International Conference on Decision and Game Theory for Security, GameSec 2021,held in October 2021. Due to COVID-19 pandemic the conference was held virtually. The 20 full papers presented were carefully reviewed and selected from 37 submissions. The papers focus on Theoretical Foundations in Equilibrium Computation; Machine Learning and Game Theory; Ransomware; Cyber-Physical Systems Security; Innovations in Attacks and Defenses. Preface Organization Contents Theoretical Foundations in Equilibrium Computation Computing Nash Equilibria in Multiplayer DAG-Structured Stochastic Games with Persistent Imperfect Information 1 Introduction 2 Imperfect-Information Naval Strategic Planning Problem 3 Algorithm 4 Procedure for Computing Degree of Nash Equilibrium Approximation 5 Experiments 6 Conclusion References Two Algorithms for Computing Exact and Approximate Nash Equilibria in Bimatrix Games 1 Introduction 1.1 Notation 1.2 Outline of the Paper 2 Preliminaries 2.1 Strategically Equivalent Games 2.2 Approximate Nash Equilibrium 3 A Fast Algorithm to Compute Strategically Equivalent Zero-Sum Games 3.1 Algorithmic Implications for Matrices in Mmn(R) 3.2 A Simple Example: Rock-Paper-Scissors 3.3 Algorithm and Simulations 3.4 Numerical Results 4 Approximate Nash Equilibrium Through an Affine Transformation 4.1 Approximate Nash Equilibrium 4.2 Algorithmic Implementation 4.3 Numerical Simulation 5 Conclusion A Some Auxiliary Results on Mmn(R) References Separable Network Games with Compact Strategy Sets 1 Introduction 2 Network Games 3 Example: Security Game 3.1 Security Game with the Tullock Function 4 Polynomial Network Games 4.1 Moment-Based Formulation 4.2 Nonnegative Polynomials Using Sums of Squares 4.3 Hierarchy of Semidefinite Relaxations 4.4 Examples 5 Conclusions References Machine Learning and Game Theory Countering Attacker Data Manipulation in Security Games 1 Introduction 2 Related Work 3 Preliminaries 3.1 Stackelberg Security Games (SSGs) 3.2 Partial Behavior Deception Model 3.3 Cognitive Hierarchy Approach 4 Finding Non-deceptive Attacker Behavior 4.1 Characterizing Deceptive Attacker's Behavior 4.2 RaBiS: Characterizing Behavior of Non-deceptive Attacker 4.3 Principled Approach for Low-Data Challenge 5 Maximin to Optimize Defender Utility 6 Experiments 7 Conclusion A Appendix A.1 Proof of Theorem 1 References Scalable Optimal Classifiers for Adversarial Settings Under Uncertainty 1 Introduction 1.1 Related Work 2 Model 2.1 Setting and Notation 2.2 Preliminary: Reduction of Dimensionality 2.3 Model Discussion 3 BNE Characterization and Computation 3.1 Equilibrium Characterization 3.2 Scalable Approximate Computation 3.3 Numerical Illustration 4 Online Learning 5 Concluding Remarks References Learning Generative Deception Strategies in Combinatorial Masking Games 1 Introduction 2 Related Work 3 Deception Through Attribute Masking 4 Computing Equilibrium Deception Strategies 4.1 Computing the Attacker's Best Response 4.2 Computing the Defender's Best Response 4.3 Computing Equilibrium Deception 4.4 Scalable Approximation of Equilibrium Deception Through Generative Adversarial Masking 5 Extension to Multiple Devices 6 Experiments 6.1 Near-Optimality of Generative Adversarial Masking 6.2 Systematic Large-Scale Experiments 7 Case Study 8 Conclusion References Network Games with Strategic Machine Learning 1 Introduction 2 Game Model 2.1 Interdependent in Decision Outcomes (Type 1) 2.2 Interdependent in Observable Features (Type 2) 3 Equilibrium Analysis–Linear Mechanisms 3.1 Type 1 Model Analysis 3.2 Type 2 Model Analysis 4 Equilibrium Analysis–Linear Threshold Mechanisms 4.1 Type 1 Model Analysis 4.2 Type 2 Model Analysis 5 Numerical Results 5.1 Type 1 Model, Linear Mechanism 5.2 Type 2 Model, Linear Mechanism 5.3 Type 1 Model, Linear Threshold Mechanism 5.4 Type 2 Model, Linear Threshold Mechanism 6 Discussion 6.1 Comparisons with Previous Works 7 Conclusion A Proof of Lemma1 B Proof of Proposition1 C Proof of Proposition2 D Proof of Lemma2 E Proof of Proposition3 F Proof of Lemma3 G Proof of Lemma4 H Proof of Lemma5 I Proof of Lemma6 J Proof of Proposition4 K Proof of Proposition5 References No Time to Lie: Bounds on the Learning Rate of a Defender for Inferring Attacker Target Preferences 1 Introduction 2 Related Work 3 Model 4 Lower Bound on the Number of Observations 5 Upper Bound on the Number of Observations 6 Simulation Results 7 Conclusion A Proof of Lemma4 B Proof of Lemma5 References When Should You Defend Your Classifier? 1 Introduction 2 Related Work 3 The Advanced Adversarial Classification Game 3.1 Adversary 3.2 Defender 3.3 Cost of Pure Strategies 3.4 Utility of Mixed Strategies 3.5 Expected Payouts for Mixed Strategies 4 Game Instantiation and Analysis 4.1 Best Response Analysis of the Adversary 4.2 Best Response Analysis of the Defender 4.3 (Fully) Mixed Nash Equilibria 4.4 Results 5 Discussion 6 Conclusion References Ransomware A Mechanism Design Approach to Solve Ransomware Dilemmas 1 Introduction 2 Related Work 3 System Model 4 Proposed Mechanisms 4.1 Ransomware-Dilemma-1 4.2 Ransomware-Dilemma-2 5 Smart-Contract Implementation 6 Conclusion References Winning the Ransomware Lottery 1 Introduction 2 Related Work 3 Probability and Lotteries 4 Paying to Play 4.1 Lowering the Value of Payments 4.2 Increasing Costs 4.3 Decreasing Payment Probability 5 Lowering the Stakes 5.1 Decreasing Attack Success 5.2 Cyber Insurance 5.3 Use of Decrypters 5.4 Off-Site Backups 5.5 Impact of Mitigations 6 Conclusion References Combating Ransomware in Internet of Things: A Games-in-Games Approach for Cross-Layer Cyber Defense and Security Investment 1 Introduction 1.1 Related Work 1.2 Organization of the Paper 2 Problem Formulation 2.1 Basic Settings 2.2 Multi-phase Multi-stage Game Formulation 2.3 Cyber Markov Game for Ransomware Penetration 2.4 Solution Concept 3 Ransomware Game Analysis 3.1 Risk Assessment Outcome of the Cyber Markov Game 3.2 Equilibria of the Ransomware Game 4 Sensitivity Analysis and Impact of Human Factors 4.1 Impact of the Security Budget 4.2 Impact of Human Factors and Prospect Theory 5 Case Studies and Discussion 5.1 Model Implementation 5.2 Outcome of Cyber Markov Game 5.3 Impact of Budget 5.4 Prospect Theory 6 Conclusion A Proof in the Budget Dilemma References Cyber-physical Systems Security A Game-Theoretic Framework for Controlled Islanding in the Presence of Adversaries 1 Introduction 2 Related Work 3 Model and Preliminaries 3.1 Power System Model 3.2 Stackelberg Game 4 Problem Formulation 4.1 Adversary Model 4.2 Grid Operator Model 4.3 Interaction Model Between the Grid Operator and Adversary 5 Solution Approach 5.1 Mixed Integer Nonlinear Bi-level Optimization Formulation 5.2 Double Oracle Algorithm Based Approach 6 Numerical Evaluations 6.1 Simulation Setup 6.2 Case Study Results 7 Conclusion References Game Theoretic Hardware Trojan Testing Under Cost Considerations 1 Introduction 2 Game Theoretic Trojan Testing Under Cost Considerations 3 Game Theoretic Trojan Testing Under a Cost Budget Constraint 3.1 NE Under Sufficient Cost Budget of the Defender 3.2 NE Under Insufficient Cost Budget of the Defender 4 Numerical Results 5 Conclusion References Strategic Remote Attestation: Testbed for Internet-of-Things Devices and Stackelberg Security Game for Optimal Strategies 1 Introduction 2 Background 2.1 Software Vulnerabilities and Exploitation in IoT Devices 2.2 IoT Remote Attestation 2.3 Stackelberg Security Games 3 Testbed Design and Development 3.1 Testbed Components 3.2 Testbed Development 4 Game-Theoretic Model of Remote Attestation 4.1 Environment and Players 4.2 Strategy Spaces 4.3 Utility Functions 4.4 Solution Concept 5 Analysis of Optimal Attestation Strategies 5.1 Case 1: Single Device and Single Attestation Method 5.2 Case 2: Multiple Devices and Single Device Class 5.3 Case 3: Multiple Devices and Multiple Device Classes 6 Numerical Results 6.1 Experimental Results from the Remote Attestation Testbed 6.2 Evaluation of Game-Theoretic Model and Optimal Strategies 7 Related Work 7.1 IoT Security Testbeds 7.2 Remote Attestation 7.3 Stackelberg Security Games 8 Conclusion and Future Work References Innovations in Attacks and Defenses Bet and Attack: Incentive Compatible Collaborative Attacks Using Smart Contracts 1 Introduction 2 A Model for a CSC-Based Collaborative Attack 2.1 Blockchain Model 2.2 Threat Model 2.3 Attack Model 3 Game Theoretic Model and Analysis 3.1 Attackers Contribution 3.2 Interdependent Attackers Game (IAG) 3.3 Equilibrium Analysis 4 Exploring Incentive Compatibility 4.1 Mechanism Formulation 4.2 Incentive Compatible Property 4.3 Budget Constraint 4.4 Voluntary Participation Constraint 4.5 Fairness 5 Numerical Simulations and Discussion 6 Conclusion References Combating Informational Denial-of-Service (IDoS) Attacks: Modeling and Mitigation of Attentional Human Vulnerability 1 Introduction 1.1 Related Works 1.2 Notations and Organization of the Paper 2 System Modeling of Informational Denial-of-Service Attacks 2.1 High-Level Abstraction and Motivating Example 2.2 Sequential Arrivals of Alerts Triggered by Feints and Real Attacks 2.3 Manual Inspection and Attention Management 3 Semi-Markov Process Model for Performance Evaluation 3.1 Consolidated State and Consolidated Cost 3.2 Long-Term Risk Measures for IDoS Attacks 3.3 Inter-arrival Time with Independent PDF 3.4 Data-Driven Assessment 4 Numerical Experiments and Analysis 4.1 Value Iteration and TD Learning 4.2 Severity Level and Aggregated Risk Without Attention Management 4.3 Severity Level and Aggregated Risk with Attention Management 5 Conclusion References Password Strength Signaling: A Counter-Intuitive Defense Against Password Cracking 1 Introduction 1.1 Contributions 1.2 Related Work 2 Preliminaries 3 Strength Signaling and Password Storage 3.1 Account Creation and Signaling 3.2 Generating Signals 3.3 Delayed Signaling 4 Adversary Model 4.1 Adversary Utility: No Signaling 4.2 Optimal Attacker Strategy: No Signaling 5 Information Signaling as a Stackelberg Game 5.1 Attacker Utility 5.2 Optimal Attacker Strategy 5.3 Optimal Signaling Strategy 6 Experimental Design 7 Empirical Analysis 7.1 Password Strength Signaling Against Offline Attacks 7.2 Password Strength Signaling Against Online Attacks 7.3 Discussion 8 Conclusions References Evaluating Attacker Risk Behavior in an Internet of Things Ecosystem 1 Introduction 2 Background 3 Methods 4 Defining the Game 5 Evaluation 6 Conclusion References Paying Firms to Share Cyber Threat Intelligence 1 Introduction 2 Model 2.1 Cyber Threat Intelligence Sharing Game 2.2 Research Questions 3 Results 3.1 Minimal Payments 3.2 The Importance of Differentiated Incentives 3.3 Case Study: A Cascading Failure on Regular Graphs 4 Conclusion References Author Index