2022/03/23 by Christopher Harth-Kitzerow, Harth-Kitzerow, Christopher, Gonzalo Munilla Garrido +1
Computer Science · #Advanced Neural Network Applications #Adversarial Robustness in Machine Learning #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #IoT and Edge/Fog Computing
paper · pdf · doi:10.48550/arxiv.2203.12347
openalex publication_date 2022/03/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
An edge computing marketplace could enable IoT devices (Outsourcers) to outsource computation to any participating node (Contractors) in their proximity. In return, these nodes receive a reward for providing computation resources. In this work, we propose a scheme that verifies the integrity of arbitrary deterministic functions and is resistant to both dishonest Outsourcers and Contractors who try to maximize their expected payoff. We tested our verification scheme with state-of-the-art pre-trained Convolutional Neural Network models designed for object detection. On all devices, our verification scheme causes less than 1ms computational overhead and a negligible network bandwidth overhead of at most 84 bytes per frame. Our implementation can also perform our verification scheme's tasks parallel to the object detection to eliminate any latency overhead. Compared to other proposed verification schemes, our scheme resists a comprehensive set of protocol violations without sacrificing performance.