Collective Intelligence

I like to think of collective intelligence as the nice things that can happen when you put together entities that interact, either by competing or by collaborating, in an as simple a way as possible.

I started developing an interest by co-investigating some aspects of collective intelligence in congestion games; therein agents compete for resources and it turns out that some simple protocols, where agents act selfishly, can end up in a situation where everybody is more or less happy.

• D. Kalles, A. Kaporis, V. Mperoukli, and A. Chatzinouskas. ”Discovery of Emergent Sorting behavior using Swarm Intelligence and Grid-enabled Genetic Algorithms”, Biologically-Inspired Techniques for Knowledge Discovery and Data Mining, S. Alam, G. Dobbie, Y. Koh, and S. ur Rehman (eds), IGI Global, pp. 249-271, 2014.

• D. Kalles, V. Mperoukli, A. Papandreadis. ''Emerge-Sort: Swarm Intelligence Sorting'', Panhellenic Conference on Artificial Intelligence, Lamia, Greece, 2012.

• D. Kalles, A.C. Kaporis and P.G. Spirakis. “Myopic distributed protocols for singleton and independent-resource congestion games”, 7th International Workshop on Experimental Algorithms, Provincetown, MA, May-June 2008.

We then looked into how some very simple operators can be very useful for sorting, without even saying which way you need to sort! That's Emerge-Sort.

The collective intelligence theme has been a central one in our multi-agent systems work with an emphasis on gaming.