New Developments in Multilinear Maps: Zeroizing Attacks, Defenses, and Post-Zeroizing Obfuscation
summary
Shows how a family of CLT graded encodings yields a zeroizing attack that multiplies selected low‑level encodings into top‑level zeros, assembles an integer matrix W from zero tests, recovers CRT‑component ratios by eigenvalue computation and uses those ratios to factor the public modulus x0. Defenses change zero testing to work modulo an independently chosen large prime n, search for a multiplier pzt via a short‑vector lattice step (LLL‑style) exploiting CRT structure, and ships a working CLT15 prototype.
AI-written summary · about
chapters (12)
- 0:00Start
- 6:18A brief history of GES candidates candidate using ideal lattices
- 14:14[CLT13]: overview
- 18:50[CLT13]: zero-test
- 20:57The zero test in action
- 28:56The [CHLRS15] attack
- 45:04no low-level zeros
- 45:47Extension #3: matrix encodings
- 47:32How could we thwart the Cheon et al. attack?
- 49:14An approach that didn't work out
- 52:48[CLT15] in a nutshell
- 54:06CLT encodings over Z Level- encoding
related talks
01h 07mLattices and Encrypted Computation
44mBilinear Pairings-based Zero-Knowledge Proofs
37mUsing Zero to Attack Zero-Knowledge Proof (ZKP) PLONK
35mEfficient Zero Knowledge Protocols: The Modular Approach
01h 06mOn the Assumptions Used for Obfuscation
52mDiscrete Log based Zero-Knowledge Proofs
22mScalable Zero-Knowledge Protocols From Vector-OLE
40mPost-Quantum
