Dan Boneh Research Group

Stanford University

Can cryptographic proofs scale to verify large model computations without revealing model parameters? The group develops succinct zero-knowledge and folding-based SNARK constructions to reduce proof size and verification cost for large computations. Researchers design homomorphic and lattice-based encryption schemes and residue-number-system techniques to enable arithmetic on encrypted large integers. They prototype verifiable computation and authenticated-data systems for scalable image and web security, combining IOPs and accumulators to enable integrity checks at scale. The lab evaluates web, mobile, and blockchain security, producing protocols for privacy-preserving airdrops and accountable threshold signatures.