Error-Correcting Codes That Enable Efficient SNARKs
summary
Claim: building SNARKs from linear error-correcting codes yields very fast provers (operations limited to field add/mul and hashes) and plausible post‑quantum security, but proof size hinges on code distance and Merkle paths. Tools used: IOPPs/IOPs of proximity, Fiat–Shamir, Merkle trees, polynomial commitments, and code switching (e.g., Blaze) between RAA and Reed–Solomon. Concrete findings: prover cost ≈ encoding time, Reed–Solomon gives best proximity gaps, and RAA needs rows ≥ 2^20 to reach distance ≈0.19.
AI-written summary · about
chapters (14)
- 0:00Welcome
- 0:43SNARKs from error-correcting codes
- 2:12Error-correcting codes
- 5:27Talk outline
- 7:36Interactive proofs and PCPs
- 9:08Interactive oracle proofs
- 11:42Interactive oracle proofs of proximity
- 16:16Non-interactive IOPPs
- 20:33SNARK costs
- 24:05SNARK-friendly code classes
- 36:38SNARK-friendly codes
- 46:31Code switchable codes
- 53:34Prover vs verifier trade-offs
- 57:24Open questions
related talks
01h 09mProofs - Practice 1
39mPrivacy and Evolution of SNARKs
01h 08mTutorial on Proof Systems
23mSonic: Nearly Trustless SNARK Setups
01h 15mProofs - Theory
52mFaster, Smaller, Weaker, Darker
18mA Selection of Pairing Based zk-SNARKs
01h 09mNoisy Connections: A Survey of Interactive Coding and Its Borders with Other Topics
