Report on the Advanced Course on Artificial Intelligence (ACAI-01) incorporating The Third European Agent Systems Summer School

Kurt Driessens

The ACAI 2001 Summer School took place between July 2nd and July 13th 2001 at the Czech Technical University in Prague in the Czech Republic. The Co-Chairs of the event were Michael Luck (University of Southampton, UK), Robert Trappl (ÖFAI, Austria), and Vladimir Marik and Olga Stepankova (Czech Technical University in Prague, CZ).

The Summer School ran over two weeks including two workshops (Adaptability and Embodiment using MAS (AEMAS 2001) and Engineering Societies in the Agents' World (ESAW 2001)) and more than 30 invited speakers, including leading researchers in the field such as Michael J. Wooldridge of the University of Liverpool, UK, and Milind Tambe of the University of Southern California, USA.

Three speakers were supported by ILPnet2:

  1. Daniel Kudenko, who talked about full multi-agent learning as being the meeting point between Machine Learning research (normally only focussing on disembodied ML or single-agent learning) and Agents Systems research. He gave an overview of current agent-oriented learning techniques spanning inductive learning, reinforcement learning and discovery.

  2. Dimitar Kazakov focussed on the application of Inductive Logic Programming to Multi-Agent-Systems, discussing some of the differences between attribute value learning and ILP, and presented the York MA Environment as a framework in which multi-agent learning experiments can be carried out.
    Slides from Kudenko and Kazakov talk

  3. Kurt Driessens' presentation went more in to detail, focussing on Relational Reinforcement Learning, a learning technique that combines reinforcement learning with relational learning. Through the use of some sample applications some of the benefits of the technique were illustrated.
    Slides from Driessens talk

Other topics discussed at the Summer School included:

Further information on ACAI 2001 can be found at