The Multi-source Probabilistic Inference (MUPI) research group studies statistical machine learning and artificial intelligence. We develop new methods and algorithms for coping with uncertainty in artificial intelligence, focusing in particular on approximate Bayesian inference of probabilistic programs. We also solve interesting practical problems across multiple application fields, developing machine learning techniques in particular for setups with limited amount of training examples.
The group operates at the Department of Computer Science, and belongs to two broader research institutes, Helsinki Institute for Information Technology HIIT and Finnish Center for Artificial Intelligence (FCAI).