Tutorial: Security
summary
Covers precise definitions (simulation/ideal world, DP-style leakage, covert/abort variants) and a running Two‑Server Model example: computing a DP histogram without sending raw data. Compares concrete protocols—PRIO (seed expansion, one‑hot vectors, heavy D cost), OPRF/anonymous histograms (deterministic encodings + dummy injection), and distributed point functions/Poplar (log‑D client keys, O(N) server work) —and flags the hard practical tradeoffs: unbalanced helper costs and real risks with confidential VMs (attestation, rollback/state, side channels).
AI-written summary · about
chapters (8)
- 0:00Secure computation introduction
- 1:16Defining secure computation and trust
- 10:58Two-server model and threats
- 13:56Simulation-based security definitions
- 29:32Publicly verifiable covert security
- 36:14Secure computation protocols and systems
- 56:31Confidential VMs and limits
- 1:02:00Hardware vs cryptographic approaches
related talks
17mPreview: Security and Privacy
15mNecessary Conditions in Multi-Server Differential Privacy
38mHow to read a PL-security paper
37mDemystifying Privacy Preserving Computing
20mPantomime: Constructive Leakage Proofs via Simulation
25mHardening Confidential Federated Computations against Side-Channel Attacks
38mDemystifying Privacy Preserving Computing
32mBeyond Software: Hardware-Based Security