A Text-Embedding-based Approach to Measure Patent-to-Patent Technological Similarity * – Workflow, Code, and Applications – Daniel S. Hain † φ , Roman Jurowetzki φ , Tobias Buchmann ψ , and Patrick Wolf ψ φ AI:Growth Lab, Aalborg University Business School, Denmark ψ Centre for Solar Energy and Hydrogen Research Baden-Württemberg (ZSW) Abstract: This paper describes an efficiently scalable approach to measure tech- nological similarity between patents by combining embedding techniques from natural language processing with nearest-neighbor approximation. Using this methodology we are able to compute existing similarities between all patents, which in turn enables us to represent the whole patent universe as a technological network. We validate both technological signature and similarity in various ways, and demonstrate at the case of electric vehicle technologies their usefulness to measure knowledge flows, map techno- logical change, and create patent quality indicators. Thereby the paper contributes to the growing literature on text-based indicators for patent analysis. We provide thorough documentations of the method, including all code, indicators, and intermediate outputs at https://github.com/ANONYMEOUS_FOR_REVIEW). Keywords: Technological similarity; patent data; natural-language processing; tech- nology network; patent landscaping; patent quality * All code necessary to recreate our workflow, indicator creation, and analysis is freely available at https://github.com/daniel-hain/patent_embedding_research. We further provide an inter- active visualization platform www.gpxp.org, which allows exploration and insight creation of all developed indicators and their geographical distribution, similarity networks between countries and technology, and their development over time. All data is also available there for download and own analysis. We hope thereby to spur further research and method development based on semantic indicators of technological development. Financial support for ZSW’s research by BMBF Kopernikus ENavi (FKZ:03SFK4W0) † Corresponding author: dsh@business.aau.dk 1