Berkeley RDI and Polyhedra unveiled the zkML system, providing new opportunities to enhance trust and transparency in artificial intelligence without exposing sensitive data.
zkML Technology: Working Principles
The zkML technology is based on applying zero-knowledge proofs (ZKP) to machine learning. ZKP is a cryptographic technique allowing one party to prove the truth of a statement without revealing the associated data. This solution helps address trust issues in 'black box' systems, which often lack transparency.
From Theory to Practice
The concept of zkML was first introduced in 2020 by Jiaheng Zhang and Berkeley researchers Yupeng Zhang and Dawn Song. At the time, zkML was theoretical due to high computational demands. Thanks to new advancements in zero-knowledge technology, such as Polyhedra's Expander proof system, zkML can now be applied in real-world scenarios.
Future Applications of zkML
zkML has the potential to transform how AI systems manage privacy and accountability. It facilitates data origin verification, ensures the authenticity and traceability of training data, and allows for the validation of the model training process. Polyhedra envisions zkML's role in combining AI and blockchain, supporting decentralized AI ecosystems and privacy-focused solutions.
zkML promises to transform trust approaches in AI by ensuring safety and privacy. Polyhedra and Berkeley RDI plan to enhance zkML capabilities, making it accessible for developers with minimal cryptography knowledge.
