BNAIC 2012 Maastricht

Jos Uiterwijk, Nico Roos, Mark Winands

Department of Knowledge Engineering

MaastrichtUniversity

{uiterwijk,roos,m.winands}@maastrichtuniversity.nl

On October 25-26, 2012, the 24^th Benelux Conference on Artificial Intelligence (BNAIC) was held in Maastricht, The Netherlands. It was organised by the Department of Knowledge Engineering of Maastricht University under the auspices of the Benelux Association for Artificial Intelligence (BNVKI) and the DutchResearchSchool for Information and Knowledge Systems (SIKS).

The conference was a lively mixture of 18 oral presentations sessions in 3 parallel tracks plus a poster and demo session plus 2 keynote lectures. The first day had sessions on Evolutionary Algorithms, Knowledge Representation, Logic I, Intelligent Agents I, Cognitive Modelling, Speech & Image Processing, Intelligent Agents II, Knowledge Discovery & Data Mining, and Planning and Scheduling. In these sessions 33 papers were presented by one of their authors. These presentations were well visited and in general of high quality.

The scientific part of the first day was concluded by a keynote lecture by Dr. Chris Welty of the IBMT.J.WatsonResearchCenter. Dr. Welty gave a talk titled Knowledge is not the Destination: IBM’s Watson and the evolving role of semantic technology. In this talk he started with the traditional vision of AI, where understanding flowed from perception through language to knowledge. It had always been envisioned that this understanding would be in some precise and unambiguous knowledge representation, and that all meaning processing would happen in this representation. This is the root of all semantic technology today. However, over time, the failure of the AI community to achieve this end-to-end vision made many, especially those in NLP, question the endpoint, in other words, to doubt the value of semantic technology. Dr. Welty next showed that it was the vision, not the technology, that deserved to be doubted: semantic technology has significant value in accomplishing tasks that require understanding, but it is not the endpoint. He underlined his ideas with an insight in the working of Watson, the computer program that defeated the two best human players in the famous American show quiz Jeopardy! in 2011.

The first day then continued in a more informal nature, in line with the Bourgondic atmosphere of Maastricht. First we had a reception at the beautiful city hall, offered by the municipality of Maastricht. Next, the conference dinner took place at the building of Thiessen Wijnkoopers BV, one of the oldest wine houses in The Netherlands (founded in 1740). After a short (in fact, too short) tour through the wine cellars of this historic building we enjoyed an exquisite diner, accompanied with a selection of very good wines. The owner revealed us that in fact the wines were not chosen to match the dinner, but the dinner was chosen to match the wines! After dinner, most of us (for those still not tired) stayed the rest of the evening in a relaxed way with a lot of chatting, laughing (and wine drinking).

The next day started early (especially for those who stayed late the previous day) with 29 presentations in sessions on Knowledge Presentation II / Agent Programming Languages, Logic II, AI for Games and Entertainment, Intelligent Agents III, Classification & Verification, Machine Learning, Semantic Web, Natural Language Processing, and Reinforcement Learning. Again, the majority of the presentations was highly interesting and of good quality. Also, during the break a session with 7 poster presentations and 5 demonstrations could be visited. Moreover, during one of the last sessions two “surprise” presentations were given, namely by two award winners. The first one was by the winner of the KION 2012 Award (the prize is awarded annually by the joint Artificial Intelligence degree programs in The Netherlands for the best master thesis), Egbert van der Wal, for his master thesis Object Grasping with the NAO. In his award lecture, Egbert gave a clear overview of his thesis research. In that a NAO robot learned to grasp small objects, but contrary to the traditional way robots learn such task by imitation, he used reinforcement learning techniques. Results showed that robots learned their tasks considerably better than by imitation only. The second award lecture was by the recipient of the newly established SNN Machine Learning Award for a Dutch or Belgian researcher or group of researchers, both public and private, who has/have achieved an important result in the field of Machine Learning, either scientifically or in the realization of an innovative application. The first of this annual award was give to Dr. Laurens van der Maaten for his work on visualizing data in multiple maps. The presentation by Laurens convincingly showed the audience that he was a deserved winner of this prize.

Again, the last scientific program of the day was concluded by a keynote lecture. This day this was delivered by Dr. Georgios Yannakakis of IT University of Copenhagen (though he recently moved to the University of Malta). In his lecture with the challenging title Game AI Is Dead – Long Live Game AI! he gave an overview of past and present in computer-games research. Dr. Yannakakis made a plea for redefining the term “game AI”. Traditionally, the tasks associated with game AI were revolved around non-player character (NPC) behavior at different levels of control, varying from navigation and path finding to decision making. Commercial-standard games developed over the last 15 years and current game productions, however, suggest that the traditional challenges of game AI have been well addressed via the use of sophisticated AI approaches, not necessarily following or inspired by advances in academic practices. The marginal penetration of traditional academic game AI methods in industrial productions has been mainly due to the lack of constructive communication between academia and industry in the early days of academic game AI, and the inability of academic game AI to propose methods that would significantly advance existing development processes or provide scalable solutions to real world problems. Recently, however, there has been a shift of research focus as the current plethora of AI uses in games is breaking the non-player character AI tradition. A number of those alternative AI uses have already shown a significant potential for the design of better games. Dr. Yannakakis presented two such key game AI research areas that are currently reshaping the research roadmap in the game AI field and evidently put the game AI term under a new perspective. These game AI flagships include the computational modeling of player experience and the procedural generation of content.

The closing session of the conference consisted mainly of the awarding of several prizes. The best paper-award, sponsored by SNN, was given to Ulle Endriss for his paper Automated Analysis of Social Choice Problems: Approval Elections with Small Fields of Candidates. The best-demo award, sponsored by SKBS, was given to Michel Klein, Nataliya Mogles, and Arlette van Wissen for their demonstration of eMate – Stimulating Behavior Change via Mobile Phone. The next two awards recipients were already revealed in a paper session for those who were present (and mentioned above), but in the closing they formally received their prizes, namely Egbert van der Wal, receiving the KION best-master-thesis award, and Laurens van der Maaten, being the recipient of the prestigious SNN Machine Learning Award.

The conference proceedings contain 33 full A-papers (new contributions), 36 extended abstracts of B-papers (papers already published elsewhere) and 5 abstracts of C-papers (demonstrations). They are published under editorship of Jos W.H.M. Uiterwijk, Nico Roos, and Mark H.M. Winands, ISSN 1568-7805, and online available at the BNAIC 2012 website (http://www.unimaas.nl/bnaic2012/). A few hard copies of the proceedings are still available for € 40 including postage via one of us (via email to one of the addresses given above).

The conference was concluded with thanking many persons, repeated here, namely the keynote speakers, Dr. Chris Welty and Dr. Georgios Yannakakis, for their inspiring contributions, the program committee (see the BNAIC 2012 website for the full list of names), the conference team for their local organizational support, the main sponsor IBM, and all other sponsors (NWO, SIKS, D-CIS Lab, Morpheus, Unilogic, SNN, SKBS, the municipality of Maastricht, Maastricht Congres Bureau, and Maastricht Booking Service) for their financial support. We think BNAIC 2012 was a great success and we already wish the team of the BNAIC 2013 conference of the Technical University Delft success in organizing the jubilee 25^th edition of BNAIC at November 7-8, 2013 in Delft, The Netherlands.