The goal of the SPIKE group is to develop data and knowledge-driven solutions for safe, robust and human interpretable AI. We investigate novel frameworks, algorithms, and scalable systems for the automated acquisition and adaptation of structured and probabilistic knowledge from (raw) data. Solutions can be used whether the knowledge is an integrated part of a complex system, or a model of complex dynamic system behaviour. We lead the field of Symbolic Machine Learning and advance the AI landscape in a number of areas groupped under three main themes: Neuro-symbolic AI, Symbolic AI and practical application to domains such as Healthcare, Decision making, Pervasive and Ubiquitous Systems, Distributed Systems, Security and Privacy.
Combining symbolic knowledge-driven AI with machine learning approaches to get "the best of both worlds".
Generating logic programs to explain symbolic data in a robust and interpretable fashion.