Linked e-resources
Details
Table of Contents
Intro; Preface; Organization; Abstracts of Keynote Talks; The Insecurity of Machine Learning: Problems and Solutions; Electronic Voting: A Journey to Verifiability and Vote Privacy; Cryptocurrencies and Distributed Consensus: Hype and Science; Contents
Part I; Contents
Part II; Machine Learning; Privacy-Enhanced Machine Learning with Functional Encryption; 1 Introduction; 2 Functional Encryption Libraries; 2.1 Implemented Schemes; 3 Implementation of Cryptographic Primitives; 3.1 Pairing Schemes; 3.2 Lattice Schemes; 3.3 ABE Schemes; 4 Benchmarks; 4.1 Inner-Product Schemes
4.2 Decentralized Inner-Product Scheme4.3 Quadratic Scheme; 5 Privacy-Friendly Prediction of Cardiovascular Diseases; 6 London Underground Anonymous Heatmap; 7 Neural Networks on Encrypted MNIST Dataset; 8 Conclusions and Future Work; References; Towards Secure and Efficient Outsourcing of Machine Learning Classification; 1 Introduction; 2 Related Work; 3 Problem Statement; 3.1 Background on Decision Trees; 3.2 System Architecture; 3.3 Threat Model; 4 Design of Secure and Efficient Outsourcing of Decision Tree Based Classification; 4.1 Design Overview; 4.2 Protocol; 4.3 Security Guarantees
5 Experiments5.1 Setup; 5.2 Evaluation; 6 Conclusion; References; Confidential Boosting with Random Linear Classifiers for Outsourced User-Generated Data; 1 Introduction; 1.1 Scope of Work and Contributions; 2 Preliminary; 3 Framework; 3.1 SecureBoost Learning Protocol; 3.2 Security Model; 4 Construction with HE and GC; 4.1 Technical Detail; 5 Construction with SecSh and GC; 5.1 Technical Detail; 6 Cost Analysis; 7 Security Analysis; 7.1 Implication of Revealing It to CSP; 8 Experiments; 8.1 Effectiveness of RLC Boosting; 8.2 Cost Distribution; 8.3 Comparing with Other Methods
8.4 Effect of Releasing It9 Related Work; 10 Conclusion; A Appendix; A.1 Boosting Algorithm; A.2 Confidential Decision Stump Learning; A.3 Cloud and CSP Cost Breakdown and Scaling; References; BDPL: A Boundary Differentially Private Layer Against Machine Learning Model Extraction Attacks; 1 Introduction; 2 Preliminaries; 2.1 Supervised Machine Learning Model; 2.2 Model Extraction with only Labels; 3 Problem Definition; 3.1 Motivation and Threat Model; 3.2 Boundary-Sensitive Zone; 3.3 Boundary Differential Privacy; 4 Boundary Differentially Private Layer; 4.1 Identifying Sensitive Queries
4.2 Perturbation Algorithm: Boundary Randomized Response4.3 Summary; 5 Experiments; 5.1 Setup; 5.2 Overall Evaluation; 5.3 BDPL vs. Uniform Perturbation; 5.4 Impact of and; 6 Related Works; 7 Conclusion and Future Work; References; Information Leakage; The Leakage-Resilience Dilemma; 1 Introduction; 2 Randomization Granularity; 2.1 Virtual-Memory Randomization; 2.2 Physical-Memory Randomization; 3 Threat Model; 4 Relative ROP Attacks; 4.1 Partial Pointer Overwriting; 4.2 RelROP Chaining; 5 RelROP Prevalence Analysis; 5.1 Analysis-Tool Architecture; 5.2 Analysis of Real-World Binaries
Part I; Contents
Part II; Machine Learning; Privacy-Enhanced Machine Learning with Functional Encryption; 1 Introduction; 2 Functional Encryption Libraries; 2.1 Implemented Schemes; 3 Implementation of Cryptographic Primitives; 3.1 Pairing Schemes; 3.2 Lattice Schemes; 3.3 ABE Schemes; 4 Benchmarks; 4.1 Inner-Product Schemes
4.2 Decentralized Inner-Product Scheme4.3 Quadratic Scheme; 5 Privacy-Friendly Prediction of Cardiovascular Diseases; 6 London Underground Anonymous Heatmap; 7 Neural Networks on Encrypted MNIST Dataset; 8 Conclusions and Future Work; References; Towards Secure and Efficient Outsourcing of Machine Learning Classification; 1 Introduction; 2 Related Work; 3 Problem Statement; 3.1 Background on Decision Trees; 3.2 System Architecture; 3.3 Threat Model; 4 Design of Secure and Efficient Outsourcing of Decision Tree Based Classification; 4.1 Design Overview; 4.2 Protocol; 4.3 Security Guarantees
5 Experiments5.1 Setup; 5.2 Evaluation; 6 Conclusion; References; Confidential Boosting with Random Linear Classifiers for Outsourced User-Generated Data; 1 Introduction; 1.1 Scope of Work and Contributions; 2 Preliminary; 3 Framework; 3.1 SecureBoost Learning Protocol; 3.2 Security Model; 4 Construction with HE and GC; 4.1 Technical Detail; 5 Construction with SecSh and GC; 5.1 Technical Detail; 6 Cost Analysis; 7 Security Analysis; 7.1 Implication of Revealing It to CSP; 8 Experiments; 8.1 Effectiveness of RLC Boosting; 8.2 Cost Distribution; 8.3 Comparing with Other Methods
8.4 Effect of Releasing It9 Related Work; 10 Conclusion; A Appendix; A.1 Boosting Algorithm; A.2 Confidential Decision Stump Learning; A.3 Cloud and CSP Cost Breakdown and Scaling; References; BDPL: A Boundary Differentially Private Layer Against Machine Learning Model Extraction Attacks; 1 Introduction; 2 Preliminaries; 2.1 Supervised Machine Learning Model; 2.2 Model Extraction with only Labels; 3 Problem Definition; 3.1 Motivation and Threat Model; 3.2 Boundary-Sensitive Zone; 3.3 Boundary Differential Privacy; 4 Boundary Differentially Private Layer; 4.1 Identifying Sensitive Queries
4.2 Perturbation Algorithm: Boundary Randomized Response4.3 Summary; 5 Experiments; 5.1 Setup; 5.2 Overall Evaluation; 5.3 BDPL vs. Uniform Perturbation; 5.4 Impact of and; 6 Related Works; 7 Conclusion and Future Work; References; Information Leakage; The Leakage-Resilience Dilemma; 1 Introduction; 2 Randomization Granularity; 2.1 Virtual-Memory Randomization; 2.2 Physical-Memory Randomization; 3 Threat Model; 4 Relative ROP Attacks; 4.1 Partial Pointer Overwriting; 4.2 RelROP Chaining; 5 RelROP Prevalence Analysis; 5.1 Analysis-Tool Architecture; 5.2 Analysis of Real-World Binaries