A dynamic approach for identifying technological breakthroughs with an application in solar photovoltaics

Bixuan Sun, Sergey Kolesnikov, Anna Goldstein, Gabriel Chan

Research output: Contribution to journalArticlepeer-review

10 Scopus citations

Abstract

This paper presents a conceptual framework for understanding technological breakthroughs and a novel empirical approach for investigating potential breakthrough inventions in the patent record. We define technological breakthroughs as inventions that are initially novel to a technological field but become increasingly relevant for describing the development of the field over time. We operationalize these notions of novelty and relevance by applying topic modeling to a corpus of the full text of patents in a technological field. The identified topics define a “technological space,” from which we develop continuous measures of a patent's novelty and relevance to the mainstream trajectory of technological development in this space over time. Our method allows us to identify potential breakthrough inventions that may have driven changes to the rate and direction of technological development in the focal field. We apply the method to silicon solar photovoltaic patents granted in the United States between 1977 and 1996, generating a list of 98 patents representing potential breakthroughs that can be subsequently validated with other approaches. This method can help researchers identify sources and patterns of technological breakthroughs to inform research and development policy.

Original languageEnglish (US)
Article number120534
JournalTechnological Forecasting and Social Change
Volume165
DOIs
StatePublished - Apr 2021

Bibliographical note

Publisher Copyright:
© 2020

Keywords

  • Innovation
  • Solar photovoltaics
  • Technological breakthrough
  • Technological space
  • Technological trajectory
  • Topic modeling

Fingerprint

Dive into the research topics of 'A dynamic approach for identifying technological breakthroughs with an application in solar photovoltaics'. Together they form a unique fingerprint.

Cite this