Tracing the evolution of service robotics: Insights from a topic modeling approach

  • Authors:

    Ott, I., Savin, I., Konop, C.

  • Source:

    Technological Forecasting and Social Change, 2022, Vol. 174

  • Date: January 2022
  • We propose a machine learning approach to identify service robotics technologies within the knowledge space of robotic patents.

    Apart from analyzing this particular area of knowledge for the first time using machine learning methods, our paper provides two important methodological novelties:

    (i) We propose a new matching method to distinguish topics belonging to service robotics using descriptions of SR fields provided by the international Federation of Robotics.

    (ii) We construct a graph of topic interrelations showing which topics appear more often together thus distinguishing between more "general purpose" topics and those which appear more isolated

    (typically specific application fields). In doing that, we distinguish significant edges by building a null model and conducting a Monte Carlo experiment to assess how likely the edge is to be observed given the structure of the data.

    The paper offers a series of findings regarding dynamics of the topics over time, the share and position of SR topics in the overall space of knowledge, and the implications of our findings for policy