Using the Internet of Everything for Machine-Learning-Based Computer System Design and Optimization

Machine learning (ML) has been widely used in computer system development and optimization levels, boosting computer design and optimization improvement. With the increase of computer system design complexity and the boost of the demand for software and application optimization, the difficulties of computer system problem solving are increasing, stimulating the development of ML methods. ML models represent computer system design and development approaches; combining these with the Internet of Everything (IoE) cannot be ignored. For ML models, all kinds of computer system sample data are the fundamental prerequisite for quantifying the computer system characteristics; the connection (internet) of all computer devices is critical for predicting computer system behaviors. The IoE enables more computer devices and ML models to be connected, extending how computer systems can be well developed. The IoE will affect the applications between computer system development and various ML models more deeply by connecting people, data, and machines. This work provides an overview of the history of the IoE and ML-based computer system development. It shows the interaction between ML models and computer system design and optimization. It also emphasizes the people-to-people (P2P), people-to-machine (P2M), and machine-to-machine (M2M) in the ML-based computer system design and optimization. It discusses the interpretability of ML-based computer system development.