Decentralizing Machine Learning Operations using Web3 for IoT Platforms

Abstract

Combining machine learning with IoT offers new and exciting possibilities for service innovation. A core challenge for such systems is model training and pipeline monitoring/management while not revealing sensitive data or introducing new attack points. Innovative solutions that reveal network structure and topology are often in conflict with security/privacy design principles. In this paper, we consider the system communication between sensors, edge, and the cloud that does not reveal the origin source or destination of the data. A core contribution is how to apply Machine Learning Operations (MLOps) to decentralized architectures that are not under the direct control of a service provider. The aim is to decouple the Machine Learning (ML) solution, data platform, and sensors while avoiding service level degradation. We utilize an unlinkable end-to-end encrypted asynchronous communication protocol called Whisper that is based on the Ethereum blockchain to achieve sender and receiver anonymity. The Whisper protocol provides a level of darkness that maintains anonymity in communication.