schematic of tribal formation and decayDavid Hales

I do research in the tribler.org team at Delft University of Technology in the Parallel and Distributed Systems Group. My background is Computer Science and Artificial Intelligence but I have spent a lot of time with Sociologists, Philosophers and even lapsed Economists doing simulations. Why do I mention this? I suppose I am trying to indicate that I don't approach Computers in a "traditional way" so I'm not occupied with traditional problems and techniques.

Complex networked "always on" information systems are increasingly displaying properties comparable to biological and social entities. Crude mechanical analogies and patterns in software design are holding back progress - we are not building "engines" anymore. A more suitable approach is to consider such systems as "societies" and program them with ideas derived from social theory - particularly new ideas coming from Computational Sociology, Evolutionary Economics and the like.

I'm developing approaches for growing reliable, decentralized systems - specifically peer-to-peer systems. The aim is to produce self-configuring, self-repairing and self-organising societies of  peer nodes that execute on user machines. I want to harness desirable emergent properties of massive (millions of nodes) systems under conditions of high flux with nodes constantly entering and leaving the system, and possibly behaving maliciously. In these kinds of "anarchic environments" system behaviour needs to emerge "bottom-up" from the behaviour of individuals following their own rules. Imposition of central control is not an option and we can't rely solely on the "hidden hands" of classical economic theory. So a synthesis is required between new kinds of social theory applicable to the artificial domain and distributed self-organising systems programming. Such a synthesis could transform, not just technology but, human societies in profound ways. My aim is to encourage transformations that are for the good of everyone - not just a technologically or financially privileged few.

Group Selection in Information Systems

My current approach uses "tag" and network "re-wiring" mechanisms to produce systems in which groups or "tribes" of agents or peer nodes emerge dynamically within information systems to service user needs. These "virtual tribes" have some of the characteristics of human groupings or "tribes" such as, dynamic formation and dissolution over time, cooperation and specialisation within the tribe, and sometimes, antagonistic behaviour between tribes. Understanding how to harness these virtual tribal dynamics to produce robust and productive systems is what I am concentrating on.  This tribal approach has produced promising results in a number of  application domains but more work needs to be done.

In the News:

new scientist magazine cover
al
P2Pnet.net logo
atlas magazine front cover
slashdot logo
New Scientist Agent Link p2pNet
ATLAS (Turkish)
Slashdot


Recent talk:

At this workshop of European Conference on Complex Systems (Oct. 2007)
I discuss how to remove the need for banks and transform the world money system
using P2P technology and social networks, amoung other things ;-)
Click on picture to watch talk:


Emergent Networks as Distributed Reputation System

Brief History

I received a BSc in Computer Science from Aston University and an MSc and a PhD from Essex University (under the supervision of Jim Doran). Until July 2001 I was a senior research officer in the area of Machine Learning and Data Mining at Essex University (with Paul Scott). From Sept 2001 until Jan 2002 I was a visiting scholar at the Italian National Research Council (CNR) in Rome working with Rosaria Conte's group. From Jan 2002 until Aug. 2003 I was a research fellow at the Centre for Policy Modelling (CPM) at Manchester Metropolitan University (MMU) working with Scott Moss and Bruce Edmonds. Between Jan. 2004 and Dec. 2007 I was based in the Dept. of Computer Science at the University of Bologna, Italy working in Ozalp Babaoglu's group on the DELIS project. Since Jan 2008 I have been here in Delft, Holland, working in the tribler.org team on the P2P-Next project.

Other Activities

I am currently a member of the editorial board of the Journal of Artificial Societies and Social Simulation (JASSS). I was involved in the formation of the new European Social Simulation Association (ESSA) and am currently an elected member of the management committee. I have been, and am, a member of the programme or organising committees for numerous agent-based simulation and self-organising applications related workshops and conferences (recent ones include: ESSA04, MAMABS04, ESOA04, ABS04, M2M04, SIC05, ESOA05, ESSA05, MASTA05, SOAS05, ECE05, ESOA06, WCSS06, MABS06, EUC06, SBIA06, CSS-TW1, ECCS06, P2P06, EPOS06, M2M07, P2P07, SASO07, ESSA07, MABS07, ICPP07, MASTA07, MA4CS07, EEMMAS07, P2P-NVE07, EUMAS2007, SOPDS2008, WCSS2008, ESSA2008, AAMAS2008, MABS2008, SASO2008).

For contact details see the bottom of this page.


Journal Articles

Journal Features / Reviews / Commentary

Edited Volumes and Special Issues

Conferences and Book Chapters

Technical Reports and Discussion Papers

Posters

Talks


Contact Details

Dr David Hales
Delft University of Technology
Department EWI/PDS/Tribler
PO Box 5031
2600 GA Delft
The Netherlands
Email: dave (at) davidhales.com
Web: www.davidhales.com
Tel: +31 (0)1527 85804 (office)
Tel: +44 (0) 208 133 4466 (skype-in with voicemail)
Fax: +31 (0)1527 86632 (please put for atten. of David Hales)




Last Updated: June 2008
eXTReMe Tracker