BACK TO ARCHIVE

Verifiable Inference for Edge Compute

Paper
Artificial Intelligence
15-10-2025

ABSTRACT

This paper introduces novel cryptographic protocols for verifying the integrity of machine learning inference performed on untrusted edge computing nodes. We develop lightweight proof systems that enable clients to verify that edge servers correctly executed ML models without revealing sensitive model parameters or input data.

TECHNICAL FOCUSEdge Computing, Verifiable Computation, ML Security, Cryptographic Proofs
RESEARCH CATEGORYArtificial Intelligence
PRIMARY OUTPUTPDF

RELATED RESEARCH