In the DIVA research group we not only make use of the state-of-the-art Deep Learning technologies, but we actively investigate their nature and behaviour to better understand these complex architectures.
Deep learning is a fast moving field which has seen a large rise in popularity in the last decade. In our research group we ride this wave of progress without forgetting the traditional field of machine learning where it comes from. We strive for knowledge and understanding and we are driven by challenge.
Embracing these new algorithms, we conduct foundamental research on artificial neural networks: we investigate their learning behavior, develop new models and training techniques and then we measure their performances under different setting and scenarios.
Our most prominent fields of application are Historical Document Image Analysis and Digital Pathology. Moreover, we do make active research in Reinforcement Learning, especially related to Artificial Intelligence for games.