A recent study published in Engineering introduces an innovative framework that integrates machine learning (ML) and blockchain technology (BT) to improve computational security in engineering applications. This framework, called Machine Learning on Blockchain (MLOB), seeks to overcome the shortcomings of current ML-BT integration methods, which mainly concentrate on data security while neglecting computational security. ML is extensively utilized in engineering for solving complex problems, providing high accuracy and efficiency. Nevertheless, it is susceptible to security threats like data tampering and logic corruption. BT, known for its decentralization, transparency, and immutability, has been investigated as a means to protect engineering data. Despite this, traditional ML workflows remain exposed to off-chain vulnerabilities, as ML models are generally executed outside the blockchain environment. The MLOB framework addresses this by embedding both data and computation within the blockchain, running them as smart contracts and securing execution logs. It comprises four key components: ML acquisition, where an ML model is trained for a designated task; ML conversion, which modifies the trained model for deployment on the blockchain; ML safe loading, which safeguards the security of data and model transfer; and consensus-based ML model execution, ensuring the correctness and safety of computational processes. To demonstrate MLOB’s effectiveness, the researchers built a prototype and applied it to monitor indoor construction progress. They evaluated the framework against three baseline methods and two recent ML-BT integrated approaches.

Findings revealed that MLOB considerably enhanced security, successfully thwarting six predefined attack scenarios. It preserved high accuracy, with only a minimal 0. 001 difference in mean intersection over union (MIoU) compared to the top-performing baseline. While efficiency was slightly reduced, showing a 0. 231-second increase in latency relative to the fastest baseline, its overall performance still aligned with industrial practice demands. Additionally, the MLOB framework carries managerial significance. It motivates organizations to pursue innovation by adopting cutting-edge technologies, fostering more competitive engineering processes. It also mitigates risks linked to data and logic security, which helps optimize resource allocation and bolster economic resilience. Nonetheless, the framework has limitations, including restricted support for latency-sensitive applications and the absence of a user-friendly interface. Future work will aim to enhance efficiency and develop a more accessible interface to improve usability and broaden MLOB’s application in engineering computations.