Epistemic AI (E-pi) is a research project funded by the European Union under its Horizon 2020 initiative, in particular the Future Emerging Technologies (FET) scheme, which is designed to foster blue-sky thinking and bring about paradigm changes.
This is a four-year (48 months) project that started on 1 March 2021, with a budget of around 3 million euros and three international partners (all universities): Oxford Brookes University (United Kingdom); Katholieke Universiteit Leuven (KU Leuven, Belgium); and Delft University of Technology.
Epistemic AI breaks entirely with the current state of artificial intelligence. This includes the most high-profile ongoing efforts such as continual learning (making learning a life-long endeavour); multi-task learning (distilling knowledge from many tasks to solve a different problem); or meta-learning (learning to learn). All of these are still firmly rooted in AI’s conventional principles, which means that they fail to recognise the foundational issue that the discipline still has with the representation of uncertain knowledge.
Epistemic AI’s overall objective is to create a new paradigm for a next-generation artificial intelligence, providing worst-case guarantees on its predictions using proper modelling of real-world uncertainties. First, a new mathematical framework for optimisation under epistemic uncertainty will be formulated, superseding existing probabilistic approaches. The new optimisation framework will lay the foundations for the creation of new ‘epistemic’ learning paradigms. We will focus, in particular, on some of the most important areas of machine learning: unsupervised learning, supervised learning and reinforcement learning.