A new study published in Engineering presents a novel framework that combines machine learning (ML) and blockchain technology (BT) to enhance computational security in engineering. The framework, named Machine Learning on Blockchain (MLOB), aims to address the limitations of existing ML-BT integration solutions that primarily focus on data security while overlooking computational security.
ML has been widely used in engineering to solve complex problems, offering high accuracy and efficiency. However, it faces security threats such as data tampering and logic corruption. BT, with its features of decentralization, transparency, and immutability, has been explored to safeguard engineering data. But the traditional ML process remains vulnerable to off-chain risks as ML models are often executed outside the blockchain.
The MLOB framework places both data and the computational process on the blockchain, executing them as smart contracts and protecting execution records. It consists of four core components: ML acquisition, where an ML model is trained for a specific task; ML conversion, which adapts the trained model for blockchain deployment; ML safe loading, ensuring the security of data and model transfer; and consensus-based ML model execution, guaranteeing the safety and correctness of computations.
To illustrate the effectiveness of the MLOB framework, the researchers developed a prototype and applied it to an indoor construction progress monitoring task. They compared the MLOB framework with three baselines and two recent ML-BT integrated approaches. The results showed that the MLOB framework significantly enhanced security, successfully defending against six designed attack scenarios. It maintained a high level of accuracy, with only a 0.001 difference in the mean intersection over union (MIoU) metric compared to the best baseline method. Although it had a slightly compromised efficiency, with a 0.231 second latency increase compared to the most efficient baseline, the overall performance met the requirements of industrial practice.
The MLOB framework also has managerial implications. It encourages organizations to innovate by integrating advanced technologies, which can lead to more competitive engineering operations. It also helps mitigate risks associated with data and logic security, optimizing resource allocation and enhancing economic resilience.
However, the framework has some limitations. It has limited support for latency-sensitive scenarios and lacks a user-friendly interface. Future research will focus on optimizing its efficiency and designing a more accessible user interface to further improve its usability and expand its application in engineering computing.
The paper "Machine Learning on Blockchain (MLOB): A New Paradigm for Computational Security in Engineering," authored by Zhiming Dong, Weisheng Lu. Full text of the open access paper: https://doi.org/10.1016/j.eng.2024.11.026
