Indistinguishability Obfuscation from Well-Founded Assumptions
summary
Constructs iO by reducing it to a structured seed PRG built from two assumptions: sparse LPN and an NC0 Boolean PRG with stretch n^{1+τ} (τ arbitrarily small, e.g. 0.001). Public seed is LPN samples as public seed that hide the input; a short private seed stores precomputed low-degree monomials and compressed correction data via low-rank matrix bucketing to fix about M·λ/L^δ errors. Evaluation uses bilinear (degree‑2) operations; security loses a 1/λ factor but admits amplification.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:53Learning Parity with Noise
- 3:58Boolean PRGs
- 6:50Structured PRGs
- 10:23Implausibility Results
- 13:30Constructing the Model
- 27:49Finding a Short Seed
- 28:29Preprocessing
- 31:29Bad output bits
- 35:24One naive approach
- 54:18Questions
related talks
01h 08mExpedition to Obfustopia: Indistinguishability Obfuscation from Well-Studied Assumptions to New...
58mObfuscation 2
42mObfuscation IV: A New Assumption enabling iO
01h 26mObfuscation 1
01h 04mTutorial on obfuscation
46mIndistinguishability Obfuscation from Bilinear Maps and LPN Variants
01h 12mApplications of Obfuscation I
52mIndistinguishability Obfuscation via Mathematical Proofs of Equivalence
