Efficient and Affordable Zero-Knowledge Proofs: Trillion Gates and Beyond
Xiao Wang
Abstract:
Zero-knowledge (ZK) proofs with an optimal memory footprint have attracted a lot of attention because such protocols can easily prove very large computation with a small memory requirement. In this talk, I will talk about some recent progress on concretely efficient ZK protocols and their applications in this setting. Our protocols are very cheap computationally and can prove large statements like ResNet inference with ease; on the other hand, it is designated-verifier, and the proof size is often linear to the circuit size. The unique features of our protocol make it highly affordable. With the cost of one Caffe Latte (Grande size, tax excluded), we could prove 3.65 trillion AND gates at a speed of 0.2 microseconds (~1-bit communication) per AND gate.