Artificial Intelligence Security and Privacy

This two-volume set LNCS 14509-14510, constitutes the refereed proceedings of the First International Conference on Artificial Intelligence Security and Privacy, AIS&P 2023, held in Guangzhou, China, during December 35, 2023. The 40 regular papers and 23 workshop papers presented in this two-volume set were carefully reviewed and selected from 115 submissions. Topics of interest include, e.g., attacks and defence on AI systems; adversarial learning; privacy-preserving data mining; differential privacy; trustworthy AI; AI fairness; AI interpretability; cryptography for AI; security applications.

Inhalt

Fine-grained Searchable Encryption Scheme.- Fine-grained Authorized Secure Deduplication with Dynamic Policy.- Deep Multi-Image Hiding with Random Key.- Member Inference Attacks in Federated Contrastive Learning.- A network traffic anomaly detection method based on shapelet and KNN.- DFaP: Data Filtering and Purification Against Backdoor Attacks.- A Survey of Privacy Preserving Subgraph Matching Method.- The Analysis of Schnorr Multi-Signatures and the Application to AI.- Active Defense against Image Steganography.- Strict Differentially Private Support Vector Machines with Dimensionality Reduction.- Converging Blockchain and Deep Learning in UAV Network Defense Strategy: Ensuring Data Security During Flight.- Towards Heterogeneous Federated Learning: Analysis, Solutions, and Future Directions.- From Passive Defense to Proactive Defence: Strategies and Technologies.- Research on Surface Defect Detection System of Chip Inductors Based on Machine Vision.- Multimodal fatigue detectionin drivers via physiological and visual signals.- Protecting Bilateral Privacy in Machine Learning-as-a-Service: A Differential Privacy Based Defense.- FedCMK: An Efficient Privacy-Preserving Federated Learning Framework.- An embedded cost learning framework based on cumulative gradient.- An Assurance Case Practice of AI-enabled Systems on Maritime Inspection.- Research and Implementation of EXFAT File System Reconstruction Algorithm Based on Cluster Size Assumption and Computational Verification.- A Verifiable Dynamic Multi-Secret Sharing Obfuscation Scheme Applied to Data LakeHouse.- DZIP: A Data Deduplication-Compatible Enhanced Version of Gzip.- Efficient Wildcard Searchable Symmetric Encryption with Forward and Backward Security.- Adversarial Attacks against Object Detection in Remote Sensing Images.- Hardware Implementation and Optimization of Critical Modules of SM9 Digital Signature Algorithm.- Post-quantum Dropout-resilient Aggregation for Federated Learning via Lattice-basedPRF.- Practical and Privacy-Preserving Decision Tree Evaluation with One Round Communication.- IoT-Inspired Education 4.0 Framework for Higher Education and Industry Needs.- Multi-agent Reinforcement Learning Based User-Centric Demand Response with Non-Intrusive Load Monitoring.- Decision Poisson: From universal gravitation to offline reinforcement learning.- SSL-ABD:An Adversarial Defense MethodAgainst Backdoor Attacks in Self-supervised Learning.- Personalized Differential Privacy in the Shuffle Model.- MKD: Mutual Knowledge Distillation for Membership Privacy Protection.- Fuzzing Drone Control System Configurations Based on Quality-Diversity Enhanced Genetic Algorithm.- KEP: Keystroke Evoked Potential for EEG-based User Authentication.- Verifiable Secure Aggregation Protocol under Federated Learning.- Electronic voting privacy protection scheme based on double signature in Consortium Blockchain.- Securing 5G Positioning via Zero Trust Architecture.- Email Reading Behavior-informed Machine Learning Model to Predict Phishing Susceptibility. <p