Scientific Report

DECOI: Design of Collective Intelligence 2008

February 23-27, 2009

The third International workshop on the design of collective intelligence (DECOI) aimed at providing a research and discussion platform within the relatively new and fragmented field of large scale, autonomous and adaptive systems. DECOI covers a spectrum of topics within this field to provide participants with state-of-the-art knowledge about methods and techniques (evolutionary computing, artificial life models, self organising systems) as well as related research domains (sociology, economics, biology). The workshop is setup as a combination of lectures and projects on which participants work during the week. Especially this latter aspects a fruitful environment for the development new ideas and future collaborations and overcomes the problem of participants absorbing information only passively.

DECOI 2009 attracted 39 participants from 10 (European) countries. This amount could have been larger, but the organization committee specifically decided to keep the group of participants manageable and therefore closed registration early. Since the emphasis of DECOI is on ‘learning by doing’, the main audience for the workshop where Master/Graduate and PhD students. During the registration procedure, applicants were asked to provide a short biography and motivation on which selection of participants was based.

This year DECOI presented a broad variety of speakers covering topics like crowd simulation, pervasive computing, self-organising and multi-agent systems, insect based computing and complex structural engineering. The lectures were given by renowned experts from all over Europe and gave a good overview of the state-of-the art within their respective research topics. Anders Johanson (ETH Zurich, Switzerland) showed how crowd simulation is used to analyse catastrophes occurring during the yearly Hadj in Mekka. Martin Middendorf (University of Leipzig, Germany) focussed on swarm inspired algorithms solving resource and network related problems. Franco Zambonelli (University of Modena and Reggio Emilla, Italy) covered ambient/pervasive computing and a framework to embed communicating intelligent sensors/agents throughout is environment. Giovanna Di Marzo Serugendo (University of London, UK) provided a general framework to understand self organizing systems and finally Jeroen Coenders (Arup, Amsterdam) showed the state-of-the-art in applied research and focussed on how complex adaptive systems/methodologies are currently used in structural engineering (e.g. the ‘Birds nest’ Olympic stadium in Beijing).

As already mentioned, the field of large scale, autonomous and adaptive systems is fairly fragmented into a multitude of research topics and disciplines. To address this challenge, the organizers opened the conference with a lecture that grounded many of these topics into a wider framework. Since self organising systems typically show characteristics like robustness and adaptation, they are in the storefront of demand driven research since in face of the current crises society demands systems to become more resilient towards perturbations (financial crisis, climate crisis, etc.).

Before the workshop, each of the invited speakers was asked to define a specific task/project that the participants could work on during the week. To overcome the problem of speakers just attending DECOI for a few hours for their presentation, the organisation committee specifically selected speakers that guaranteed to be present during most of the week in order to supervise the projects and be present during the final presentation. Although finally not all invited speakers were able to fulfil this prerequisite, 2 of the 5 speakers managed to stay until the final project presentation, while 2 others were attending DECOI multiple days to supervise the project progression.

Apart from the lectures and projects a number of other initiatives were developed to ensure maximum knowledge exchange between participants. These included a poster-session, various social events (drinks, dinners). All in all this resulted in a positive reaction from both participants as well as invited speakers. At the end of DECOI each participant filled in an evaluation form, and the overall response was overwhelmingly positive. The provided working environment, facilities, informal atmosphere and level of the presentations were regarded positively.

A final indicator of success of such an event is the formation of new collaborations extending into the future. At the time this report is written we cannot yet assess the impact DECOI had in this respect. DECOI developed a website (www.decoi2009.collectivae.net) as a platform on which the participants are stimulated to publish DECOI-related collaborations. During the week many initiatives were discussed to organize lectures, classes, projects and write papers.

Finally, we would like to add that the organization of DECOI became a breeze because of the extensive support from the Lorentz centre. This includes the handling of applicants prior to DECOI, publicity, daily affairs, accommodation and facilities. Without this support we are confident that DECOI wouldn’t have been the success that it seems to be.

Martijn Schut (Vrije Universiteit, Amsterdam)

William Veerbeek (Unesco-IHE, Delft)

Konrad Diwold (University of Leipzig, Leipzig)

Virginia Dignum (Utrecht University, Utrecht)