Max Welling specialises in learning systems and studies their application in the analysis of large-scale datasets. These days, unprecedented quantities of data are being generated not only by research science but also by public bodies, businesses and society in general. Buried in this flow of raw data is information that can tapped to make predictions. For example, an automated analysis of an email can be used to predict whether the email is actually spam. An analysis of a user’s interaction with his smartphone can be used to predict his particular interests at that time.
Welling plans to investigate how the intelligent application of principles of computing science (such as distributed computing) might be used to scale up sound, mathematically-founded statistical methods into very large models and datasets. At the same time, he will be exploring how the insights that neuroscience and cognitive science have provided into human learning could be utilised in computer-directed learning. Other questions that interest Welling include: How can machines be designed to engage in continuous learning and to continuously adapt the complexity of their internal models to new information (analogous to humans)? And what sorts of statistical procedures would offer scientists and businesses the best possible support towards designing models and simulations?
In addition to his professorship at the UvA, Welling has also held an appointment at the University of California Irvine (USA) since 2003. Previously he worked at Utrecht University (where he also obtained his doctorate), the California Institute of Technology (USA), the University of Toronto (Canada), University College London (UK) and Radboud University Nijmegen. Welling has received various grants in the United States, including the NSF Career Award. In 2010, Welling was presented with the ECCV Koenderink Prize and, in 2012, with the ICML’s Best Paper Award. He is associate editor in chief van IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI).