ZK-FPE

Blockchain: Research and Applications

Background

Model ownership verification requires disclosing sensitive weights or architecture details, creating a privacy-versus-accountability dilemma.

Content

We combine blockchain and zero-knowledge proofs to build verifiable model ownership attribution. Fingerprint vectors enable scalable transfer via vector addition.

Core Contribution

ZK-FPE achieves privacy-preserving ownership verification: the verifier can confirm fingerprint presence without accessing model internals. ZK-VOT extends this to on-chain/off-chain mutual trust.

Experimental Result

The framework supports scalable, efficient fingerprint transfer and establishes verifiable trust between on-chain contracts and off-chain computation.

Key Numbers

• Privacy-preserving ownership verification
• Scalable fingerprint transfer via vector addition
• On-chain/off-chain mutual trust (ZK-VOT)