Applied MPC
summary
Three concrete contributions: a protocol that runs PageRank over partitioned bank graphs using Secure PageRank via MPC with fixed-point arithmetic and an iteration-stopping rule whose revealed iteration count leaks one sample from a geometric distribution tied to the second eigenvalue lambda_2; a formal compiler model that maps high-level functions to protocol-tailored classes called M-Circuit compilation (STEM‑circuits), captures gadgets like Beaver triples and defines perfect/statistical insecurity from trace differences; and an implementation of Oblivious TLS with secret-shared DH using EC Diffie‑Hellman, preprocessed masked points, threshold EdDSA, ~1 s handshake over LAN and ~3 MB/s AES-GCM online throughput.
AI-written summary · about
chapters (11)
- 0:00Intro
- 0:20Analytic methods for fraud detection
- 4:10Privacy enhancing technologies
- 5:53Linear secret sharing
- 7:22Early termination of power method
- 9:28Conclusion
- 12:51Traditional non-Cryptographic Compilation
- 23:45Multiparty Computation (MPC)
- 27:32Oblivious TLS
- 29:06Handshake Layer: Key Exchange in MPC
- 33:20Estimation of Performance
related talks
01h 08mSecure Computation 1
01h 03mSecure Computation 2
29mTowards Scalable Constant-Round MPC from Minimal Assumptions via Round Collapsing
45mPrice of Active Security in Multiparty Computation
58mSecure Multiparty Computation I
01h 01mMPC with Silent Preprocessing via Pseuodorandom Correlation Generators
01h 03mClosing Keynote Day 2
26mEfficient Multiparty Protocols via Log-Depth Threshold Formulae