Development of a smart patent recommendation system with natural language processing capabilities

Artificial Intelligence (AI) and machine learning are increasingly adopted in diverse areas such as medicine, manufacturing, finance, transportation, retailing, and supply chain management to enhance productional and operational efficiency and smart decision making. This research develops an intelligent patent recommendation system using AI techniques and non-supervised machine learning (M/L) with natural language processing (NLP) for technology mining. Technology e-discovery for specific smart machines and manufacturing systems, such as sensors, controllers, and cyber physical systems (CPS), are used as case examples to demonstrate the prototype patent recommender. The recommendation system, trained for other domains, can be configured as a generic patent recommender. Collaborative filtering and content-based M/L and NLP approaches are adopted to implement the patent recommender with self-learning patent search capabilities. The proposed patent recommender can provide predictions for future research and development, avoiding intellectual property infringement, and provide proactive protection for market advances.