Here is my proposal to create rather evolve bio-intelligence.
Evolution of Bio-intelligence
My proposal is to use bacteria as a form of crude neural networks.. such as to play a game of tic-tac-toe or perhaps classifying images or playing chess…
This will be a form of unsupervised learning where:
-we would culture a bacteria colony/ or many bacterial colonies simultaneously
-encode inputs as signals/chemicals understood by bacteria
-encode outputs as signals/chemicals excreted by colony
-provide food if the output of colony is correct that is the move made by bacteria is correct one
-provide toxins if output is incorrect or the move made is illegal
-provide a combination of both (proportionally) if the output is fuzzy
-if need arises introduce mutagens into the colony to accelerate evolution
The rest of work will be done by the evolution or by the colony itself.
This way the best chess playing bacterial colony will evolve, the rest will die. And This way we can create real intelligence in the real-world and which is not just a simulation as in case of artificial neural networks.
End of proposal
Resources: A similar work by Erach has been published in AI newsletter. Click here to access it.
UPDATE (7th Jan, 2006)
From my discussions at Google groups, I have gathered few interesting points:
1. “” My skeptical mind says that evolution would work, but it
is too slow. However, suddenly immunology popped into my mind, and the way that it can respond to antigens (usually proteins) that is has never seen before and respond. B cells and T cells use a set of DNA segments that they can combine in MILLIONS of different sequences to generate receptors for proteins they have never seen before. If the protein is actually encountered, it signals that this new combination is actually useful and this cell then proliferates and multiplies. Perhaps incorporating such
strategies (somatic hypermutation, clonal selection) into bacteria would be
necessary for it to rapidly evolve the necessary molecular network that
could respond with the proper signals (in this case to play chess) “” says Greg
2. Quorum sensing: We know that bacteria are able to communicate between themselves and also they are also able to change the expression of certain genes according to exterior environment.