Assessment of patent quality - challenges for patent intelligence stakeholders using digital technologies

Abstract

Assessing patent quality is a challenging task in patent management, especially in patent intelligence, for two reasons: firstly, because of the ambiguity of the definition of quality and secondly, because of the handling of a vast amount of patent data, particularly in relation to the qualitative comparative analyses traditionally performed by patent practitioners. This dissertation addresses both challenges and explores how to assess patent quality in patent intelligence using digital technologies by considering patent intelligence stakeholders and their perspectives. After providing a brief contextual background to this dissertation, five publications are presented, one of which defines patent quality and its economic, legal, and technological dimensions, and four of which introduce new methods for assessing these dimensions. For example, machine learning is used to predict the legal quality of patent applications by modeling statutory novelty according to patent law, and text-mining is used to assess the legal quality of a patent by evaluating the patentability of a patent application through semantic comparison. In this context, the dissertation addresses and discusses the challenges associated with the implementation of such digital technologies for patent quality assessment, e.g. the explainability of machine learning algorithms leading to mistrust or the challenge of heterogeneity in text-mining. Entailing several implications for management and scholarship, this dissertation extends stakeholder theory by applying it to patent quality, and provides a more comprehensive and multidimensional view by integrating different stakeholder interests.