Cyber Security Cryptography and Machine Learning: 5th International Symposium, CSCML 2021, Be'er Sheva, Israel, July 8–9, 2021, Proceedings

Preface
Organization
Contents
Programmable Bootstrapping Enables Efficient Homomorphic Inference of Deep Neural Networks
1 Introduction
2 Preliminaries
2.1 Torus and Torus Polynomials
2.2 Probability Distributions
3 Discretized TFHE
3.1 Encoding/Decoding Messages
3.2 Description
3.3 Leveled Operations
4 Programmable Bootstrapping
4.1 Blind Rotation
4.2 Look-Up Table Evaluation
5 Application to Neural Networks
5.1 Layers Without PBS
5.2 Layers with PBS
6 Experimental Results and Benchmarks
7 Conclusion
A Complexity Assumptions Over the Real Torus
B Algorithms
B.1 Blind Rotation
B.2 Sample Extraction
B.3 Key Switching
References
Adversaries Strike Hard: Adversarial Attacks Against Malware Classifiers Using Dynamic API Calls as Features
1 Introduction
2 Problem Statement
3 Adversarial Learning Background
4 Related Work
5 Design and Implementation
5.1 Data Set Collection and Features Extraction
5.2 Target BlackBox Models
5.3 Malware Evasion Using GAN (MEGAN) and MEGAN with Reduced Perturbation (MEGAN-RP)
5.4 Malware Evasion Using Reinforcement Agents
6 Evaluation Results
7 Conclusion and Future Work
References
Privacy-Preserving Coupling of Vertically-Partitioned Databases and Subsequent Training with Gradient Descent
1 Introduction
1.1 Related Work
1.2 Outline
2 Hidden Set Intersection
3 Secure Gradient Descent
3.1 Regression
3.2 Classification
3.3 Gradient Descent Approach
3.4 MPyC
4 Performance
4.1 Run-Time
4.2 Accuracy
5 Conclusions and Future Work
References
Principal Component Analysis Using CKKS Homomorphic Scheme
1 Introduction
2 Preliminaries
2.1 CKKS Homomorphic Encryption Scheme
2.2 Principal Component Analysis (PCA)
2.3 Goldschmidt's Algorithm
2.4 R2 Score
3 Vector Operations
3.1 Norm and Inversion by Norm
3.2 Ciphertext Packing
3.3 Vector Operations on Ciphertext and Sub-ciphertexts
4 Homomorphic Evaluations
4.1 Homomorhpic Goldschmidt's Algorithm
4.2 Homomorphic Power Method
4.3 Homomorphic PCA
5 Implementation Details and Results
5.1 Parameter Selection
5.2 Results
6 Conclusion and Future Work
References
DepthStAr: Deep Strange Arguments Detection
1 Introduction
2 Goals
3 Pattern Description
4 Methodology
4.1 A Formal Outline of the Algorithm
4.2 Suggested Workflow to Find Exploitable Security Weaknesses
5 Implementation
5.1 The angr Framework
5.2 Implementation Details
6 Evaluation
6.1 Rediscovery of Known Weaknesses in libcurl
6.2 Newly Detected Weaknesses
6.3 Synthetic Evaluation
7 A More General Take Away
8 Conclusion
References
Robust Multivariate Anomaly-Based Intrusion Detection System for Cyber-Physical Systems
1 Introduction
2 Threat Model
3 Proposed Methodology
3.1 Anomaly Detection Algorithm-Denoising Autoencoder (DAE)
3.2 Localization of the Attack Points
4 Experiments and Results
4.1 Dataset
4.2 Training Phase
4.3 Performance Evaluation Phase
4.4 Robustness in the Presence of Adversary During Training
5 Deployment of DAE in Real Time
6 Conclusion
References
Privacy-Preserving Password Strength Meters with FHE
1 Introduction
2 Fully Homomorphic Encryption
2.1 Privacy Preserving Search
2.2 Privacy Preserving Index Search
3 Privacy Preserving Password Strength Meters
3.1 Privacy Preserving Markov Model
3.2 Privacy Preserving PCFG Model
4 Conclusion and Future Work
References
Automatic Detection of Water Stress in Corn Using Image Processing and Deep Learning
1 Introduction
2 Proposed Approach
2.1 Dataset
2.2 Proposed Method
3 Results
4 Conclusions
References
Tortoise and Hares Consensus: The Meshcash Framework for Incentive-Compatible, Scalable Cryptocurrencies
1 Introduction
1.1 Consensus, Money, and Contracts
1.2 Permissionless Consensus via PoW
1.3 Importance of Incentive-Compatibility
1.4 Drawbacks of Leader Election
1.5 Our Contributions
1.6 Related Works
2 Informal Protocol Overview
3 Meshcash Security
3.1 Security Proof Overview
References
Game of Drones - Detecting Spying Drones Using Time Domain Analysis
1 Introduction
2 Background
2.1 Video Coding Algorithms
3 Related Work
4 Adversary Model and Proposed Detection Scheme
4.1 Detection Model
4.2 Detecting FPV Channels
5 Influence of Physical Stimulus
5.1 Lab Experiments
6 Evaluation
7 Conclusions and Future Work
References
Privacy Vulnerability of NeNDS Collaborative Filtering
1 Introduction
2 The NeNDS Algorithm
3 Privacy Attack on NeNDS
4 NeNDS Shortcomings
5 Conclusions
References
Lawful Interception in WebRTC Peer-To-Peer Communication
1 Introduction
2 Background and Related Work
2.1 Browsers' Support and Open Source WebRTC Libraries
2.2 ETSI Reference Model for Lawful Interception
2.3 Current Solutions for Intercepting VoIP Calls
3 WebRTC
3.1 Connection Initiation
3.2 Encryption
3.3 P2P Communication
3.4 Multi-party Conversations
4 The Interception Model
4.1 Signaling Services
4.2 Web Applications
5 Showcase
5.1 Signaling Services
5.2 Web Applications
5.3 LEA Management Console
6 Limitation of the Current Work
7 Conclusion
References
Hierarchical Ring Signatures Immune to Randomness Injection Attacks
1 Introduction
2 Hierarchical Signature Scheme
2.1 Preliminaries and Notation
2.2 Definition of Hierarchical-Signature Scheme
3 New Security Model
3.1 Anonymity Model
3.2 Strong Unforgeability Model
4 Modified Specific HRS Scheme
4.1 Unforgeability Analysis
4.2 Anonymity Analysis
5 Implementation
6 Conclusion
References
Theoretical Aspects of a Priori On-Line Assessment of Data Predictability in Applied Tasks
1 Introduction
2 Description and Problem Definitions
3 Metrics of Predictability: Related Work
3.1 Selection of a Predictor Based on the Model of Losses from Erroneous Predictions
4 Model and Procedure for Choosing a Predictor
5 “Ontological” Factors in Probabilistic Models of Prediction
6 Conclusion
References
Randomly Rotate Qubits, Compute and Reverse for Weak Measurements Resilient QKD and Securing Entanglement
1 Introduction
2 The Random Basis Encryption Scheme
3 Securing Entanglement
4 WM and the Random Basis CNOT QKD Scheme
References
Warped Input Gaussian Processes for Time Series Forecasting
1 Introduction
2 Preliminaries
3 Warped Input Gaussian Process Model
3.1 Model
3.2 Training
3.3 Forecasting
3.4 Modelling Seasonality
3.5 Time and Space Complexity
4 Empirical Evaluation
4.1 Synthetic Datasets
4.2 Real-World Datasets
5 Related Work
6 Conclusion
References
History Binding Signature
1 Introduction
2 Preliminaries
2.1 Verifiable Secret Sharing
2.2 Verifiable Secret Public Sharing
2.3 Verifiable Random Functions
3 History Binding Signature
4 Conditions for a Valid Signature
4.1 Unforgeability
4.2 Security
4.3 Correctness (Signing)
4.4 Correctness (Key-Revealing)
5 Conclusion and Future Work
References
Effective Enumeration of Infinitely Many Programs that Evade Formal Malware Analysis
1 Introduction
2 Foundations of Computation Theory
3 Recursive Function Theory
4 Theoretical Impossibility of a Complete formal Malware/Non-malware Program Classification
5 Discussion and Directions for Further Research
References
DNS-Morph: UDP-Based Bootstrapping Protocol for Tor
1 Introduction
1.1 Our Contribution
2 Related Work
3 Threat Model
4 Obfsproxy Design
5 DNS-Morph Design
6 DNS-Morph Reliability
6.1 Received Packets Acknowledgments
6.2 Sorting Received Packets
6.3 DNS-Morph Identifiers' Encryption and Decryption
6.4 DNS-Morph Multiple Sessions Support
7 DNS-Morph Encoded Packets
8 DNS-Morph: Security Analysis
8.1 Censor's DPI Capabilities
8.2 DNS-Morph DPI Resistance
8.3 Additional Attacks and Resistance
8.4 Active Probing and Replay Attack Resistance
8.5 Domain Names' Entropy
9 DNS-Morph Design Considerations
9.1 Choice of DNS
9.2 Choice of Base32
9.3 Query Types
9.4 Recursive DNS
10 Tests and Results
10.1 Test Setup
10.2 Client's Testing Environment
10.3 Deep Packet Inspection Tools
11 Summary
11.1 Future Works
References
Polynomial Time k-Shortest Multi-criteria Prioritized and All-Criteria-Disjoint Paths
1 Introduction and Related Work
2 Finding Prioritized Multi-criteria k-Shortest Paths in Polynomial Time
3 Prioritized Multi-criteria 2-Disjoint (Node/Edge) Shortest Paths
4 k-Disjoint All-Criteria-Shortest Paths
References
Binding BIKE Errors to a Key Pair
1 Introduction
2 Specific Proposals for BIKE
3 Practical Considerations and the BIKE Additional Implementation Package
4 Conclusion
References
Fast and Error-Free Negacyclic Integer Convolution Using Extended Fourier Transform
1 Introduction
2 Preliminaries
3 Efficient Negacyclic Convolution
3.1 Redundant Approach
3.2 Non-redundant Approach
4 Analysis of Error Propagation
4.1 Error Propagation Through FFT and FFNT
5 Implementation and Experimental Results
5.1 Benchmarking Results
5.2 Performance on Long Polynomials
5.3 Error Magnitude and Correctness on Long Polynomials
6 Conclusion
A Proof of Proposition 1
References
Efficient Secure Ridge Regression from Randomized Gaussian Elimination
1 Introduction
1.1 Approach
1.2 Roadmap
2 Preliminaries
3 Ridge Regression
4 MPC Setting
5 Solving Systems of Linear Equations
6 Secure Linear Algebra
6.1 Secure Determinant
6.2 Secure Matrix Inversion
6.3 Secure Linear Solver
7 Secure Ridge Regression
8 Performance Evaluation
9 Concluding Remarks
References
PolyDNN Polynomial Representation of NN for Communication-Less SMPC Inference
1 Introduction
2 Previous Work
3 Neural Network as Polynomial Functions in a Single Node Case
4 Multiple Layers Approximation
5 Communication-Less MPC for Polynomial Calculations
6 Distributed Communication-Less Secure Interference for Unknown DNN
7 Experiments
8 Conclusions
References
Use of Blockchain for Ensuring Data Integrity in Cloud Databases
1 Introduction
2 Background and Related Work
3 The Proposed Method
3.1 The Proposed Method Description
3.2 The Proposed Method Algorithms
3.3 The Proposed Method Potential Vulnerabilities
3.4 Correctness of the Proposed Method
3.5 The Proposed Method Attack Detection
3.6 The Proposed Method Recovery from Attack
3.7 The Proposed Method Scalability
4 Conclusions and Future Work
References
Invited Talk: The Coming AI Hackers
1 Introduction
2 Hacks and Hacking
2.1 The Ubiquity of Hacking
3 AIs Hacking Us
3.1 Artificial Intelligence and Robotics
3.2 Human-Like AIs
3.3 Robots Hacking Us
4 When AIs Become Hackers
4.1 The Explainability Problem
4.2 Reward Hacking
4.3 AIs as Natural Hackers
4.4 From Science Fiction to Reality
5 The Implications of AI Hackers
5.1 AI Hacks and Power
6 Defending Against AI Hackers
References
Turning HATE into LOVE: Compact Homomorphic Ad Hoc Threshold Encryption for Scalable MPC
1 Introduction
1.1 Our Contributions
1.2 Application: One-Server, Fault-Tolerant MPC
1.3 Related Work
2 Threshold Encryption (TE) Definitions
2.1 Threshold Encryption Syntax
2.2 Threshold Encryption Flexibility
2.3 Threshold Encryption Security
2.4 Threshold Encryption with Homomorphism
2.5 Threshold Encryption Compactness
3 Sender-Compact Ad Hoc Threshold Encryption
3.1 t-Flexibility
3.2 Reducing the Public Key Size
4 Recipient-Compact Homomorphic Ad Hoc Threshold Encryption
4.1 Building HATE from Homomorphic Encryption and Secret Sharing
4.2 Building HATE from Obfuscation
References
Fully Dynamic Password Protected Secret Sharing: Simplifying PPSS Operation and Maintenance
1 Introduction
1.1 Password Protected Secret Sharing
1.2 TOPPSS Overview
1.3 Dynamic Subset PPSS
1.4 Clientless Server Enrolling and Disenrolling in PPSS
1.5 Our Contributions
1.6 Organization
2 Background
3 Dynamic Subset TOPPSS
3.1 Overview
3.2 Specification
3.3 Performance
3.4 Security
3.5 Correctness
3.6 Robustness with Dynamic Subset TOPPSS
4 Fully Dynamic TOPPSS
4.1 Enrollment Scheme
4.2 Linear Communication Enrollment Scheme
5 Conclusion
6 Appendix: Disenrollment Scheme
References
Early Detection of In-Memory Malicious Activity Based on Run-Time Environmental Features
1 Introduction
2 Dataset
3 Pre-processing and Model Generation
4 Detection Framework
5 Experimental Evaluation
6 Conclusion
References
Software Integrity and Validation Using Cryptographic Composability and Computer Vision
1 Background
1.1 Computer Vision and Cryptographic Awareness
2 Cryptographic Awareness
2.1 Universal Composability and Real vs Ideal
3 Synthesizing Backend Communications from GUI Renderings
4 Awareness-Based Development Paradigm
4.1 Online Self-validation
5 Summary
References
Efficient Generic Arithmetic for KKW *8pt
1 Introduction
1.1 A Use Case for Balanced ZKP
1.2 Our Contribution and Outline of the Work
1.3 Intuition: MPC-in-the-Head, ch31CCS:KatKolWan18 and Our Work
2 Related Work
3 Notation
4 ch31CCS:KatKolWan18 Background
5 Adding Arithmetic to Boolean Circuits
5.1 Ring Circuits with Efficient Dot Product
5.2 Converting Between Boolean and Arithmetic
6 Our Semi-honest MPC Protocol
7 Performance Estimation
References
Trust and Verify: A Complexity-Based IoT Behavioral Enforcement Method
1 Introduction
2 Related Work
2.1 Device Identity Detection
2.2 IoT Behavior and Autonomous Techniques
2.3 Complexity and Predictability
3 Data Format and Collection
4 Device Complexity Classification
4.1 Device IP Complexity
4.2 Device Variance
4.3 Aggregate Complexity
5 Behavior
5.1 Novelty Detection
5.2 Novelty Detection Tuning Using Device Complexity
5.3 Static Hyper-Parameter
5.4 Dynamic Hyper-Parameter
5.5 Complexity-Tuned Dynamic
6 Enforcement Architecture
6.1 Enforcement
7 Results
7.1 Static Hyper-Parameter
8 Conclusions and Future Work
References
Using a Neural Network to Detect Anomalies Given an N-gram Profile
1 Introduction
2 Problem Definition
3 Anomaly Detection Methods
3.1 Static Binomial Method (SB)
3.2 EWMA
3.3 PEWMA
3.4 LSTM Anomaly Filter (LAF)
4 Application Experiment
4.1 Results from Aggregations
4.2 Slow HTTP
5 Related Work
6 Conclusion
References
Meta-X: A Technique for Reducing Communication in Geographically Distributed Computations
1 Introduction
1.1 Background on MapReduce
1.2 Motivating Examples
1.3 Problem Statement and Our Contribution
2 The System Setting
3 Meta-MapReduce
3.1 Meta-MapReduce Working
3.2 The Call Function
3.3 Meta-MapReduce for Skewed Values of the Joining Attribute
3.4 Meta-MapReduce for an Identical Location of Data and Mappers
4 Extensions of Meta-MapReduce
4.1 Incorporating Meta-MapReduce in G-Hadoop and Hierarchical MapReduce
4.2 Large Size of Joining Values
4.3 Multi-round Computation
5 Versatility of Meta-MapReduce
6 Conclusion
References
Blindly Follow: SITS CRT and FHE for DCLSMPC of DUFSM (Extended Abstract)
1 Introduction
2 Replicated State Machine Vs. CRT DFSM or DUFSM
3 Polynomial Based CRT DUFSM
4 Concluding Remarks
References
Implementing GDPR in Social Networks Using Trust and Context
1 Introduction, Background, and Related Work
2 The Trust-Based Model and GDPR Implementation
3 Experimental Evaluation
4 Conclusion
References
Author Index