A natural language processing-based approach for automating IoT search
Cheng Qian, William Grant Hatcher, Weichao Gao, Erik Balsch, Chao Lu, Wei Yu · 2023
In recent years, the world has been inundated with innumerable smart Internet of Things (IoT) devices. A number of service providers are connecting IoT devices so that a variety of smart-world services can be provided. The ever-increasing volume of IoT devices helps to connect edge entities and their parent organizations, but also leads to unnecessary data redundancies among these organizations. Thus, to better serve all organizations, and reduce these inefficiencies, it is necessary to design effective and intelligent open IoT search systems. In this chapter, natural language processing (NLP) machine learning techniques automatically transform human requests into semantic queries, enabling a computer intelligence IoT search engine, called the Automatic Coordinated QUery IoT Search Engine (ACQUISE). To achieve ACQUISE translation, a set of strategies and algorithms (baseline, static and dynamic strategies with caching replace algorithms) designed using the spaCy natural language toolkit offer the efficiency of automatic IoT query generation. In the static strategy, ACQUISE optimizes a small keyword pool that can store words related to the most used queries, while in the dynamic strategy, several caching update algorithms keep the small keyword pool updated based on received queries. We conduct extensive experiments with ACQUISE and our results confirm the efficacy of our designs where the dynamic strategy achieves 37.8% faster response than the baseline algorithm.