High-Speed zkSNARKs without Trusted Setup

on YouTube ↗25m

" summary

Claims a family of zkSNARKs that needs no trusted setup and achieves sub-linear verification and tiny proofs by combining the sum‑check protocol with extractable polynomial commitments and a reusable computation commitments preprocessing step. A compiler called Spark turns sparse multilinear matrices into dense encodings so prover work stays linear (not quadratic) and evaluation proofs remain succinct. Implementations with Hyrax, COPIS, Kairos and Cerebrus show the fastest prover among general‑purpose, shortest proofs, and measured performance on million‑constraint R1CS instances.

AI-written summary · about

" chapters (9)

1. 0:00Start
2. 1:22What is a zkSNARK?
3. 2:23Many approaches to build zkSNARKS Early theory: Short PCPs + Merkle trees (AFL591, Killan92, Micali94), ...
4. 3:03Recent focus: ZKSNARKs without trusted setup
5. 4:12Challenges with achieving succinct verification
6. 8:35Background: The sum-check protocol (LFKN90)
7. 11:00Encoding R1CS as sum-check instances
8. 17:24Our solution: SPARK
9. 18:58Spartan's techniques in a nutshell

" related talks

• 26mzk-SHARKs: Combining Succinct Verification and Public Coin SetupMadars Virza · 2nd ZKProof Workshop 2019
• 22mDIZK: A Distributed Zero Knowledge Proof SystemHoward Wu · 2nd ZKProof Workshop 2019
• 25mLegoSNARK: Modular Design and Composition of Succinct Zero Knowledge ProofsDario Fiore · 2nd ZKProof Workshop 2019
• 34mSNARKs and their Practical ApplicationsEran Tromer · Securing Computation (Simons Institute, June 2015)
• 17mSNARK Flipper. FLIP & Prove multiple instances efficientlyNikitas Paslis · ZKProof 6 in Berlin
• 25mSonic: zkSNARKs from Linear Size Universal and Updatable SRSSean Bowe · 2nd ZKProof Workshop 2019
• 44mBilinear Pairings-based Zero-Knowledge ProofsJens Groth · 2nd ZKProof Workshop 2019
• 51mWhy you shouldn't trust meKeerthana Ganesh, Shubham Patil · NDC Oslo 2022