Sign In

Communications of the ACM

ACM TechNews

Researchers Develop Protocols to Validate Integrity of ML Models

View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
Researchers Develop Protocols to Validate Integrity of ML Models

Researchers at Texas A&M University (TAMU) and the University of California, Berkeley have validated the integrity of machine learning (ML) models, using cryptographic algorithms called zero-knowledge proof protocols. The protocols are a mathematical method that enables the owner of an ML model to generate a succinct proof to demonstrate with overwhelming probability that something is true without sharing extra data.

TAMU's Yupeng Zhang and colleagues designed new zero-knowledge proof methods and optimizations engineered to transform the computations of a decision-tree model into zero-knowledge proof statements. Zhang said, "These protocols will allow the owner of a machine learning model to prove to others that the model can achieve a high accuracy on public datasets without leaking any information about the machine learning model itself."

From Texas A&M University
View Full Article


Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA


No entries found