Notes on Univariate Sumcheck
Malcom Mohamed
TL;DR
The paper addresses efficient verification of multilinear extension evaluations by coupling univariate and multivariate sumcheck techniques through a novel mlex-to-unex adaptor. It introduces a square/non-square folding strategy with generalized CRT interpolation, enabling a univariate proof system that preserves completeness and achieves tight soundness with Schwartz–Zippel checks, while maintaining linear prover time and logarithmic verifier effort. A key contribution is natural round reductions, allowing early stopping of the multivariate phase and compatibility with single-round verifiers like Aurora, reducing communication and domain overhead. The work offers practical improvements over kernel interpolation and quotient-based approaches, delivering asymptotically optimal prover time $O(2^m)$ and a streamlined, round-efficient adaptor for unex-based sumchecks. This provides a flexible, scalable framework for integrating univariate and multivariate sumchecks in SNARK-like proofs and related cryptographic protocols.
Abstract
This note describes a univariate polynomial interactive oracle proof for multilinear extension evaluation. Unlike prior protocols, (1) the verifier here is given a univariate extension oracle for the same vector of which the multilinear extension is getting evaluated and (2) the prover only has linear complexity. For these reasons, the protocol is well-suited for combining multivariate and univariate sumcheck techniques.
