Workshop on computational methods in social science



On July 1st, the Chair of Innovation and IP Policy (IIPP) of Prof. de Rassenfosse organized a workshop on “Computational methods in social science: A focus on textual content”.

The IIPP chair exploits its position at EPFL, a hub of excellence on machine learning and data science, to push the frontier of the textual analysis of patent and scientific publication data.

The workshop gathered around 30 participants from Switzerland and abroad to discuss recent developments in the field. Topics covered included the measurement of green policy uncertainty from news articles (Julien Gobeil, HEG-SIB), topic modelling of publication articles (Omar Ballester, EPFL-IIPP), name disambiguation of inventors listed on patent documents (Deyun Yin, WIPO and Orion Penner, EPFL-IIPP), measures of patent similarity (Antonin Bergeaud, Banque de France and Dominique Guellec and Caroline Paunov, OECD), fuzzy-matching firm names with machine learning (Martin Mugnier, CREST).