BaseFold:
Efficient Multilinear Polynomial Commitment Schemes from Foldable Codes
By Hadas Zeilberger, Binyi Chen, and Ben Fisch
By Hadas Zeilberger, Binyi Chen, and Ben Fisch
Description:
BaseFold is an efficient multilinear polynomial commitment scheme - a building block which enables fast Succinct Non-interactive Arguments of Knowledge (SNARK). It was the first PCS to combine two efficient paradigms: multilinear-based SNARKs and error-correcting code based SNARKs with small proofs, thereby speeding up real-world proof systems. Furthermore, it runs over any choice of finite field, and so accommodates a much wider range of applications than prior schemes. To date, it has been used to speed-up various real-world proof systems, and is a fundamental building block for several new SNARK constructions.
Projects that use BaseFold:
Industry:
Deep-Prove: Zero-knowledge proofs for AI built by Lagrange Labs (Code)
Binius: FPGA SNARKs from Irreducible: (Code, Paper)
Research:
Basefold In the List-Decoding Regime (Paper) (Reduces BaseFold's proof size)
Khatam (Paper) (Reduces BaseFold's proof size)
Recorded Past Talks:
Notes and Blog Posts About BaseFold
From LambdaClass: (link)
From SECBIT Labs: (link)