Re: Symbiotic Intelligence

From: Anders Sandberg (asa@nada.kth.se)
Date: Tue Nov 06 2001 - 15:23:20 MST


On Tue, Nov 06, 2001 at 01:43:57PM -0800, hal@finney.org wrote:
> The problem seemed somewhat artificial. It appeared that there was no
> actual collective problem-solving effort. Rather, the research simulated
> the effects if people with different information about the maze pooled
> their ideas to help find the best path. It would be interesting to see
> if the results could be replicated on a real-world problem.

While this might be less interesting than actual pooling of solution
strategies (I know I have a paper on that lying around somewhere, it
also works), it is likely a very practically relevant situation.

If the cost of cooperating on a problem is low enough, specialisation
across agents wins - pooling information individually gathered, testing
different solutions and so on produces a very broad problem solving
strategy. It doesn't work as well if the cost of cooperation is high or
if the solution is very narrow, so that there is little or no
information to gain from looking at non-correct possibilities.

-- 
-----------------------------------------------------------------------
Anders Sandberg                                      Towards Ascension!
asa@nada.kth.se                            http://www.nada.kth.se/~asa/
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y


This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 08:11:50 MST