Search papers, labs, and topics across Lattice.
This paper introduces zkComposer, a modular framework that enhances the efficiency of zero-knowledge machine learning (zkML) by decomposing proof construction into independent sub-proofs, thus enabling greater parallelism. By addressing the limitations of existing monolithic proof systems, zkComposer achieves significant reductions in prover and response times, demonstrating improvements of up to 6.84x on GPT-2 compared to previous zkML implementations. The findings underscore the potential of zkComposer to make zkML more scalable and practical for real-world applications while maintaining cryptographic integrity.
zkComposer achieves up to 6.84x faster proof generation for zero-knowledge machine learning, revolutionizing the scalability of private model inference.
Zero-knowledge machine learning (zkML) enables a server to perform verifiable inference while keeping model parameters private from the client. However, existing zkML systems incur prohibitive proof-generation costs. We observe that proof generation exhibits limited parallelism; that is, prover time does not decrease significantly as the number of threads increases. This limitation is because existing systems rely on monolithic proof computation, constructing a single proof for the entire machine learning model. We introduce zkComposer, a modular proof-construction framework that unlocks an additional dimension of parallelism, in addition to the parallelism in existing proof kernels. zkComposer decomposes the zkML proof of correct inference into independent sub-proofs, each covering a subset of the computation for inference e.g., each independent sub-proof can cover a subset of contiguous layers in the ML model. Adjacent sub-proofs are cryptographically linked through shared commitments to the activations from the boundary layer. zkComposer provides the same guarantees as the monolithic proof without requiring additional linking proofs or changes to the underlying cryptographic primitives. We implement zkComposer and evaluate it on three CNNs and GPT-2. We show that, on CNN workloads, zkComposer reduces prover time and response time by up to 3.25x relative to zkCNN [1]. On GPT-2, zkComposer reduces these times by up to 4.83x relative to zkGPT [2], when partitioning along the model layers. When partitioning across both model layers and input sequences in GPT-2, we show that zkComposer reduces prover time and response time by up to 6.84x relative to zkGPT [2].
