ECE 575 – Advanced Cryptographic Techniques with Applications to Real-World Systems
Fall 2025 · Graduate seminar, MWF
This graduate-level course covers advanced concepts in information and network security, intended for Master’s or Doctoral students interested in security research. Topics include advanced cryptographic techniques, formal security analysis methods, and their applications to security and privacy in NextG wireless/mobile networks and distributed systems, as well as data security and privacy (especially in machine-learning systems). The course combines traditional lectures with student-led paper presentations and discussions. Required textbook: Introduction to Modern Cryptography, 3rd Edition, J. Katz and Y. Lindell (Chapman & Hall/CRC, 2020).
Module 1: Introduction and Review of Information and Network Security
Week 1
- 8/25: Course Introduction
- 8/27: Review 1: Symmetric Key Cryptography (SKC)
- 8/29: Review 2: Public Key Cryptography (PKC)
Module 2: Provable Security (Security Notions)
Week 2
- 9/3: Perfect Secrecy
- 9/5: Semantic Security
Week 3 – Security Notions (Cont’d)
- 9/8: Semantic Security and PRGs
- 9/10: CPA Security
- 9/12: CCA Security
Module 3: Advanced Cryptographic Techniques
Week 4 – Hash and MAC
- 9/15: Integrity Protection and MACs
- 9/17: Hash Functions
- 9/19: Hash Function Applications
Week 5 – Public Key Cryptography
- 9/22: Preliminaries to PKC
- 9/24: Provable Public Key Schemes
- 9/26: Advanced Public Key Schemes
Week 6 – Secure Multiparty Computation (SMC)
- 9/29: Secret Sharing Schemes
- 10/1: Fully Homomorphic Encryption; Yao’s Garbled Circuit
- 10/3: Introduction to SMC and Oblivious Transfer
Week 7 – Cont’d
- 10/6: Garbled Circuit
- 10/8: Zero-Knowledge Proofs
- 10/10: Differential Privacy
Module 4: Privacy Attacks against Machine Learning and LLMs
Week 8
- 10/13: Differential Privacy (cont’d)
- 10/15: Comprehensive Privacy Analysis of Deep Learning: Passive and Active White-box Inference Attacks against Centralized and Federated Learning · (IEEE S&P 2019)
- 10/17: Systematic Evaluation of Privacy Risks of Machine Learning Models · (USENIX Security 2020)
Week 9
- 10/20: Extracting Training Data from Large Language Models · (USENIX Security 2021)
- 10/22: Analyzing Leakage of Personally Identifiable Information in Language Models · (IEEE S&P 2023)
- 10/24: Gradient Obfuscation Gives a False Sense of Security in Federated Learning · (USENIX Security 2023)
Module 5: Secure Neural Network Inference
Week 10
- 10/27: CryptoNets: Applying Neural Networks to Encrypted Data with High Throughput and Accuracy · (ICML 2016)
- 10/29: GAZELLE: A Low Latency Framework for Secure Neural Network Inference · (USENIX Security 2018)
- 10/31: CrypTFlow2: Practical 2-Party Secure Inference · (ACM CCS 2020)
Week 11
- 11/3: Cheetah: Lean and Fast Secure Two-Party Deep Neural Network Inference · (USENIX Security 2022)
- 11/5: BOLT: Privacy-Preserving, Accurate and Efficient Inference for Transformers · (IEEE S&P 2024)
- 11/7: From Individual Computation to Allied Optimization: Remodeling Privacy-Preserving Neural Inference with Function Input Tuning · (IEEE S&P 2024)
Module 6: Secure ML under Multiparty Settings (SMC)
Week 12
- 11/10: SecureML: A System for Scalable Privacy-Preserving Machine Learning · (IEEE S&P 2017)
- 11/12: ABY3: A Mixed Protocol Framework for Machine Learning · (ACM CCS 2018)
- 11/14: Cerebro: A Platform for Multi-Party Cryptographic Collaborative Learning · (USENIX Security 2021)
Module 7: Privacy-Preserving Federated Learning: Secure Aggregation
Week 13
- 11/17: Prio: Private, Robust, and Scalable Computation of Aggregate Statistics · (NSDI 2017)
- 11/19: Flamingo: Multi-Round Single-Server Secure Aggregation with Applications to Private Federated Learning · (IEEE S&P 2023)
- 11/21: RoFL: Robustness of Secure Federated Learning · (IEEE S&P 2023)
Week 14
- 11/24: ELSA: Secure Aggregation for Federated Learning with Malicious Actors · (IEEE S&P 2023)
- 11/26: ACORN: Input Validation for Secure Aggregation · (USENIX Security 2023)
- 11/28: No class (Thanksgiving)
Additional reading (not for presentation):
- Secure Single-Server Aggregation with (Poly)Logarithmic Overhead · (ACM CCS 2020)
- Leakage of Dataset Properties in Multi-Party Machine Learning · (USENIX Security 2021)
- Analyzing Inference Privacy Risks Through Gradients in Machine Learning · (ACM CCS 2024)
Module 8: Secure ML Protocols
Week 15
- 12/3: LOKI: Large-Scale Data Reconstruction Attack against Federated Learning through Model Manipulation · (IEEE S&P 2024)
- 12/5: Local and Central Differential Privacy for Robustness and Privacy in Federated Learning · (NDSS 2022)
Module 9: Final Project Presentations
Week 16 – Student project presentations
Schedule subject to change.