One competition, two rounds – in September last year, the Coleridge Initiative of New York University announced the "Rich Context Competition". Computer scientists from 20 teams and eight countries submitted letters of intent – four of them won against the competition and are now in the final round almost five months later. One of them is the team "Knowledge Technologies for the Social Sciences" at GESIS. We congratulate Wolfgang Otto, Andrea Zielinski, Behnam Ghavimi, Dimitar Dimitrov, Karam Abdulahhad, Narges Tavakolpoursaleh and Katarina Boland on this success. The final presentations of the finalist teams from KAIST, the University of Paderborn, the Allen Institute for Artificial Intelligence (AI2) (Seattle) and GESIS as well as the announcement of the winner will take place on February 15, 2019 via live stream over the Internet.
The contest initiators could not have chosen a better time to determine the winner of the contest: The worldwide Love Data Week will take place from the 11th to the 15th. One of this year's topics is "Open Data". How important it is to make data accessible to others is also clear against the background of the competition: research data is often difficult to access. Researchers and analysts working with data are often faced with the problem of finding out who has done research on what topics and with what results. As a result, good research results and data remain undiscovered while time and resources are wasted repeating the empirical work. The "Rich Context Competition" would like to counter this problem. "The competition challenged computer scientists to find ways of automating the discovery of research datasets, fields and methods behind social science research publications," said Julia Lane (New York University), one of the jury members.
The competition is about participants developing and identifying text analysis techniques and machine learning techniques to discover the relationships between data sets, researchers, publications, research methods and areas. Participants will demonstrate their skills to a panel of technical and social science experts: They have developed and further refined algorithms using various data corpses. Evaluation criteria include the accuracy of the techniques used, user-friendliness and novelty. All submitted algorithms will be made publicly accessible in the spirit of open source.