Enhancing Patent Screening with Machine Learning: Impacts on Innovation and Firm Performance

Abstract

For the United States Patent and Trademark Office (USPTO), the increasing number of patent applications often leads to inefficiencies in the USPTO's selection and approval processes. This paper employs machine learning algorithms to improve the ability to predict patent quality and assist human examiners. Combining human judgment with machine prediction can alleviate the biases and inefficiencies in the current patent selection system, providing practical policy recommendations for improving patent examination efficiency.