Abstract
The life-long monitoring and analysis for complex industrial equipment demands a continuously evolvable machine-learning platform. The machine-learning model must be quickly regenerated and updated. This demands the careful orchestration of trainers for model generation and modelets for model serving without the interruption of normal operations. This paper proposes a container-based Continuous Machine-Learning and Serving (CMS) platform. By designing out-of-the-box common architecture for trainers and modelets, it simplifies the model training and deployment process with minimal human interference. An orchestrator is proposed to manage the trainer’s execution and enables the model updating without interrupting the online operation of model serving. CMS has been deployed in a 1000 MW thermal power plant for about five months. The system running results show that the accuracy of eight models remains at a good level even when they experience major renovations. Moreover, CMS proved to be a resource-efficient, effective resource isolation and seamless model switching with little overhead.
Highlights
With the rapid advancement of Internet of Things and industrial automation systems, enterprises and industries are collecting and accumulating data with unparalleled speed and volume [1,2,3]
Compared with data in other sectors, industrial data have several distinct characteristics: firstly, the major parts of their data are produced from sensor networks and must be processed quickly [6]; secondly, use of these data is often limited within certain domains for the sake of data safety and regulations
This paper proposes a container-based computing platform to automate the workflow from model learning to deployment, namely Continuous Machine-Learning and Serving (CMS)
Summary
With the rapid advancement of Internet of Things and industrial automation systems, enterprises and industries are collecting and accumulating data with unparalleled speed and volume [1,2,3]. Several existing big data platforms, e.g., TFX [12] and IBDP [6], have already provided certain supports for both types of service This intelligent system uses data analysis technology and model-based systems to help decision-makers to improve the effectiveness of their decisions. This paper proposes a container-based computing platform to automate the workflow from model (re-) learning to deployment, namely CMS. A system component, namely orchestrator, is introduced to orchestrate the model training executions across multiple trainers It provides model deployments and model update services, while keeping the same model-serving processes uninterrupted. A container-based computing platform is constructed for continuous model learning and serving; A model management service, Orchestrator, is proposed to streamline the model updating process.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
