Francqui Chair - Prof Luc De Raedt

The artificial intelligence and machine learning revolutions

Recent breakthroughs in artificial intelligence (AI) and machine learning are revolutionizing our society. They form the core of new technologies such as self-driving cars, more accurate medical diagnosis systems and intelligent virtual assistants. The media rightly pay a lot of attention to these new development as our society cannot ignore the impact of artificial intelligence. At the same time, many questions arise about AI, sometimes about current applications, sometimes also more fundamental questions: What are the possibilities and limitations of AI? What are the implications for people and society? What does AI bring in the future? The introduction to the lecture series will formulate answers to a number of these questions, it will provide insight into what artificial intelligence is, and what we can and cannot expect. The answers are formulated from the perspective of a computer scientist and AI researcher, who not only develops AI techniques, but also follows and actively participates in the ongoing debate about AI.

Register here.

Inaugural lecture

Rector Herman Van Goethem and Professor Nick Schryvers, dean of the Faculty of Science of the University of Antwerp, are pleased to invite you to the inaugural lecture by Professor Luc De Raedt, laureate of the Francqui Chair 2019-2020.

Inaugural lecture:
The artificial intelligence and machine learning revolutions

Thursday 12 March 2020 at 5 p.m.
The laudation will be given by Professor Bart Goethals.

University of Antwerp – Campus Middelheim – Building G – G.010 – Middelheimlaan 1, Antwerp

After the lecture you are invited to the reception, register here. (event is postponed)

Lecture series

You are kindly invited for the lecture series entitled The next wave of AI: integrated learning and reasoning. A strong background in mathematics and/or algorithms and AI is recommended to follow the lectures.

The lectures can be followed online.

April 20,  Introduction to the course and motivation for learning and reasoning. [video; slides]

April 20, Introduction to Probabilistic graphical models [video; slides] (recommended reading: Chapter on Probabilistic Reasoning in the book of Russell and Norvig, AI: a modern appproach)

April 27, Advanced concept from logic for probabilistic inference [videos;slides]

May 4, Probabilistic logics and programming, and Statistical Relational Artificial Intelligence [videos;slides]

May 11,  Probabilistic logics - knowledge based model construction. [videos;slides]

May 11, Neuro Symbolic Computation -- DeepProbLog. [videos;slides - continue with previous set]

May 11, From Statistical Relational AI to Neuro Symbolic Computation -- integrating logic and neural networks. [videos;slides]


While the field of AI has originally focussed on reasoning, it is now focussing almost exclusively on learning. This is due to the recent breakthroughs in machine learning, which has contributed solutions for many problems in perception, natural language processing and computer vision. Although unprecedented accuracies are obtained for numerous prediction tasks, deep learning techniques are not a universal solution: they require vast amounts of data, produce black box models that are not explainable, provide no guarantees, and are very sensitive to the type of data they were trained on. At the same time, there has been a lot of progress in reasoning methods for planning, for scheduling, for multi-step reasoning, and for solving constraint problems.

This lecture series will address both learning and reasoning, and it will argue that integrated approaches are needed for genuine artificial intelligence. The distinction between learning and reasoning in artificial intelligence is sometimes also phrased as symbolic vs subsymbolic methods, data- vs knowledge-driven techniques, or type I vs type II systems (paraphrasing Kahneman).

The lecture series will cover the foundations of reasoning, which are based on logic and probability. In particular, it will introduce probabilistic graphical models, logical representations as well as solvers based on knowledge compilation.It will also cover the foundations of learning, where it will address techniques for learning logical, probabilistic and neural representations. But most of all the focus will be on the integration of learning and reasoning.

The lectures will devote a lot of attention to the integration of logic and probability in probabilistic databases, probabilistic programming languages and statistical relational artificial intelligence. Probabilistic programming is a recent trend in artificial intelligence and computer science that extends programming languages with probabilistic primitives and the ability to learn from data. Finally, it will also introduce the field of neuro-symbolic computation, which is an attempt to merge neural network representations with logic in order to obtain the best of both worlds.

About the Belgian Francqui Chair

Every year, the Francqui Foundation invites Belgian or European scientists to stay with Belgian universities, giving the opportunity to eminent researchers and scholars to teach a series of guest lectures at the host university and to participate in the scientific life of the institution.