I wrote an piece about QLlectives (and the P2P mentality) that has been included in the PerAda Magazine. I promised this months ago and kept putting it off but finally I wrote it. Anyhow, it can be found at:

http://www.perada-magazine.eu/view.php?source=003444-2010-12-06

In the introduction and conclusion I made reference to what I believe are the defining characteristics of the P2P approach (it's more than a technology). I believe that we're on the winning side and major changes are coming.

The PerAda-ASSYST summer school took place from the 21st-27th of September in Budapest, co-located and overlapping with the SASO conference. The workshop was co-sponsored by PerAda and ASSYST. It was structured in the following way: attendees were assigned to groups. Each group had the task of selecting and addressing a key application area in pervasive adaptation. At the end of the summer school each group produced a presentation outlining their application. The idea was not to produce completed solutions but rather to give a convincing presentation of how pervasive technology might be applied to solve an open problem by defining requirements, challenges and giving a conceptual design. During the week there were extended lectures given by several invited speakers (including Mark Jelasity and myself - both from QLectives of course) in addition to group sessions and social events. Slides from the lectures can be found here.

Franco Zambonelli did a good job of kicking-off events by outlining some interesting challenges around the idea of the "Urban Organism". The modern metropolis faces many challenges such as transport efficiency, security and provision of social or common services - such as infrastructure maintenance, garbage collection etc. It would appear that pervasive adaptable systems could provide new ways of approaching these challenges. Radical possibilities can be considered such as citizens providing their own local services outside of traditional top-down municipal arrangements (which may be inflexible, expensive and bureaucratic). Pervasive technologies could become enablers of positive cooperation by allowing, for example, trusted adults to keep an eye on children playing in a neighbourhood. Individualised transport service could be merged with public options via automated collaborative travel routing.  In general one could envisage the crowdsourcing of public services. Aside: interestingly, this kind of thing has been termed "the big society" by the current UK government. You can see Franco's excellent slides on some case study ideas and the final presentations by the groups here.

I attended the start and end of the summer school and found the atmosphere to be supportive and productive. I think it is important that students from a wide range of institutions and backgrounds from across Europe get time together to work out challenges and potential research directions. Increasingly the kind of application domains to which new technologies can be applied cut across traditional research areas and groups. It is far too easy for PhD students to become highly specialised on particular applications and approaches early on - shaped by their institutional or group affiliation - limiting their potential productive interactions with others. I believe this was the last summer school in this series but I hope other similar initiatives are organised in the future!

I had the privilege of attending the [k]NO(w)BODY workshop @ ECCS2010 in Lisbon.You can see details about it here. Why did I go? Because a number of people told me I might find it interesting and recently I've been trying to listen to people rather than judge them. I'm glad I went because in some ways I felt the workshop had a transformative effect on the the way I think. That's a big statement and I can not hope to capture it in a short text on an EU project blog (that nobody reads anyway). But I will give it a go. The workshop was run by João Fiadeiro who is a dance practitioner.  It was about improvisation. His background is dance. First let me discuss my intelligence gathering on the the way you pronounce "João". I have been told by Nazareno Andrade (one of the QLectives Delft team) that this should be pronounced like a light sabre (star wars) cutting through the air - jzooowow!. In any case it was obvious to me that João is some kind of honest person. He levelled with us at the start of the workshop on why he ended-up here doing this. It had to do with improvisation. One aspect that I picked up on was his concern that often when people are in an improvisation arena they think it is ok to do what the want. That is they act egocentrically. They do "their own thing" without thinking about the others around. There is a contradiction between the individual and the collective. Collective work must be more than just a lot of individuals "doing their own thing". I immediately thought about my work on the Prisoner's Dilemma (heck I always do that). Another aspect that João talked about was the idea that in order to produce productive collective improvisation we need to observe, think and then act. But to hold back on acting until the time is right. That is, one needs to have alternatives and timing and interaction should be via the environment. You should actually observe, listen, reflect and formulate alternatives and act. But don't act too late or too early. It is about timing. Indeed it seems there is a tension on when action is required or it is too late. Holding that tension is one of the key elements to what João calls "the method" which he has developed. Strangely and intuitively I believed I understood this method and felt it close to social contract theory or perhaps what might be called "socio-cognitive stigmergy" - that is a kind of social contract that is not agreed or stored in the minds of the agents. In any case we did improvisations (with many people from all kinds of different backgrounds) that seemed to produce collective outcomes that none of us anticipated but we all though were greater the sum of our parts. Something definitely happened. You should have been there man. I also felt like a child, like I had the excuses to play again. I realised I forgot how to play recently. Back to reality. I realised I was interacting with scientists from diverse disciplines in a way I'd never done before. I started to think that perhaps the way we normally interact through powerpoint presentation and ego orientated  "listen to me" contexts might be all wrong. Just a thought.
 

The European Conference on Complex Systems (ECCS2010) took place last month in Lisbon. A number of QLectives people attended including myself. This conference brings together major European researchers within the complex systems area and is highly interdisciplinary. It was people meeting at this conference series (which started in 2004 in Torino, Italy) that led to the COSI-ICT area which funds projects such as QLectives and other complexity related ICT projects. Consequently many colleagues from other related projects and groups were present and it was great to catch-up with them.

From my perspective ECCS2010 was everything a conference in this area should be. Given the highly interdisciplinary nature of complexity science you need a wide subject focus - which was evident. You also need to balance disciplinary approaches and techniques with an emphasis on crossover - where one discipline / technique has informed another. ECCS addresses this through a combination of organisational approaches: a) a plenary track with invited speakers who are leaders in their respective areas; b) parallel tracks grounded around problem / application domains - not disciplinary domains; c) satellite workshops which are very flexible and organised by attendees themselves reflecting their specific research domains, techniques, disciplines or projects and, last but not least, d) excellent social events, interaction spaces and activities that allow participants to interact "off stage" in a convivial and relaxed way. All this gives adequate scope for everyone to find a place within the conference but also, more importantly, for some real interactions to occur between those who would - in the normal run of academic life - not normally interact. Despite a general agreement in the academic world that interdisciplinary interactions should be encouraged I personally find it rare that quality interactions happen. I am not entirely sure why this is and I doubt there is any single reason. But it certainly has a lot to do with the lack of well organised, supportive and truly interdisciplinary venues for scientists to meet and exchange ideas. The ECCS is one of those venues and as such it is precious and the community owes a debt of gratitude to this years organisers and previous years organisers.

The full program of ECCS2010 can be found here and videos from the talks given will emerge over time here (there are already a number of talks availible). I will not attempt to summarise everything I saw, however, I would like to give a feel for the impressions I got from the event - obviously coloured by my own interests and obsessions. Firstly, certain themes emerged and one was that many of the problems confronting the globe appear to be related to complexity topics such as unintended consequences, co-evolutionary dynamics, out-of-equilbirum systems and highly inter-linked and emergent processes.

A specific subject which came up in a number of sessions - including the public panel on global problems - was the current financial crisis. What is it and how did we get here? What can (complexity) scientists do if anything? My impression was that it was widely agreed that a major issue, relevant for scientists, is that previous approaches to economic modelling - particularly in the restricted neoliberal efficient markets approach - were too narrow and too dominant. It appeared that the mainstream academic orthodoxy of economics committed to certain assumptions and methods too strongly and excluded other disciplines and approaches. This had a direct effect on policy (deregulation) and models of risk assessment (independent default probabilities and back fitting to historical data) that played a major role in creating the current mess. What then can complexity science do? It would be arrogant and wrong to say that this community has general answers. It does not. However, it would appear that we have approaches, communities and techniques that could help - at least not to make the same mistakes again.

A major ace the complexity community holds is that of inherent interdisciplinarity. Since complexity scientists come from diverse disciplines they do not share fundamental assumptions or models. Physicists, for example, approach the modelling of social systems in a very different way from Psychologists and Social Scientists. This often means that there are frequent debates about assumptions and approaches. This can be exasperating and difficult for many who may not have been schooled in the defence of such assumptions or think it is a waste of time to debate or defend them. However, those who "stick around" produce a community that is thoughtful about assumptions, aware of them and eager to debate them and perhaps even change them. In addition there is a wide range of technical experience covering different techniques, be this in agent-based modelling, statistical physics, social theory or ICT systems engineering etc. Given these factors it is possible for the community as a whole to examine a problem without a priori commitment  to particular assumptions or techniques. That through application and debate the right assumptions and techniques for the job-at-hand can be found in a highly practical way. Of course this kind of thing is easy to say and hard to do but I think these factors bode well for complexity science not falling into the trap that caught orthodox Economics.

Within this context Dirk Helbing (a partner in QLectives) presented the FuturICT Flagship idea in the plenary session which will address the financial system as a major research pillar. It produced a good set of questions from the audience evidencing all the above positive aspects of the complexity community.

I was also involved in a satellite workshop on Social Complexity of Informal Value Exchange (SCIVE2010) which looked at different aspects of existing and novel (ICT enabled) ways of exchanging value outside of formal financial institutions. Particularly interesting was the work of Roger Ballard who has done detailed anthropological work on the Hawala system of value transfer. Matts Eriksson also talked about the Ripple online system for value transfer (a kind of peer-to-peer money system). I presented an overview of the idea of an emerging "financial commons" based on social computing platforms. The slides of the talks can be found here.

I also attended the two day [k]NO(w)BODY workshop which was very interesting - a kind of art meets science event - but I will leave this for another blog post since there is a lot to say about it.
 

The videos of the presentations from the August COS-ICT workshop "Towards a Science of Socially Intelligent ICT" are now availible on the ASSYST website. I described the workshop previously like this:

The Science of Socially Intelligent ICT workshop at Imperial was pretty interesting. It was a truly interdisciplinary affair with real points of overlap. This is quite rare in my experience and on that basis the workshop was a success. I think the questions Jeff Johnson asked the speakers to consider in the their talks (see previous blog post) helped to create lines of connection and highlight also the differences in approaches. Again, that is important in such exploratory workshops. The keynote speaker Ricardo Baeza-Yeats, from Yahoo, gave a very strong talk discussing how recent research could inform design of new systems. Both Bruce Edmonds and Francesca Giardini gave talks which linked key findings from anthropology to the task of defining using social intelligence. Mario J. Gaspar de Silva made an important point about privacy. Simon Buckingham Shum presented some very nice examples of implemented systems that help scientists to track ideas. There were lots of other interesting talks and many people from the QLectives project were there. Videos of the presentations are scheduled to be put on the web soon - I will post a link when they are up. The Slides from my talk can be found here. I tied to go back to first principles to explain what I think socially intelligent systems are and why we could draw on existing theory from evolutionary game theory, anthropology and social sciences more generally and (more speculatively) apply a moral approach ("Moral Distributed Systems Engineering"?) by drawing on modern social contract theory.

 

The videos can now be found here.