Data Assessment has long used topic models to generate role profiles. In this context, a role profile is a topic that is learned from skills data. Now we have significantly expanded the type of data from which a topic model can be learned. In particular, we can learn Topic models from multimodal data like, for instance, text and images. To this end, we, together with colleagues from the Universities of Jena and Bielefeld, have adapted the method of moment that has been developed more than a hundred years ago by Karl Pearson. We will present our model and the associated learning algorithm at this year’s International Joint Conference on Artificial Intelligence (IJCAI).
Publication: J. Giesen, P. Kahlmeyer, S. Laue, M. Mitterreiter, F. Nussbaum, C. Staudt and S. Zarriess. Method of Moments for Topic Models with Mixed Discrete and Continuous Features. Proceedings of the 30th International Joint Conference on Artificial Intelligence (IJCAI).