Hi, i have to do an original research project (i have to discover something) due by the end of june. I was think i could do something on AI, but AI is very large field and I don''t know much about it. Anyone has any suggestion as to an AI topic i could research? Digital Radiation

Try researching the field of neural networks. They are pretty cool actually, and are gonna shape the next generation of AI.

Look in the main articles and find the one about neural nets

=====
Khaos
=====

Edited by - khaos on October 16, 2000 10:25:51 PM

I have yet to choose my dissertation for my MSc in AI (has to be behaviour based, no knowledge representation or GOFAI) but I''m likely to be going for something in Neural Networks,I''m heading in the direction of evolving asynchronous networks containing simulated hormones in a capture the flag style team testbed to see if, by the ability of individuals to express random symbols (random initially then see if they evolve), teams can increase their abilities through simple communication.

May be a bit much for 4 months work but what the Hell.

Mike

You could also do something on FSM ...although I think that it is not really something you should do for a project (is just good to know ;-) ). Otherwhise I would recommend Neural Networks too.
Or Fuzzy Logic, if you are good enough (jus kidding).


MARC

Thanx for the help guys, i''ll begin doing some background research on neural networks.

BTW, i should mention i''m a junior in high school and i''m currently taking pre-calculus. Does neural nets require me to know calculus? In the xgames3d.com article it says...

"the true nature of neural nets is extremely mathematical and understanding and proving the theorems that govern them takes Calculus, Probability Theory, and Combinatorial Analysis not to mention Physiology and Neurology."

Any other suggestions, comments, tips, etc. are welcome.


Digital Radiation

In terms of maths the most important areas I''ve found for simple nets require the ideas of summation of input and application of functions (simple things you will already have covered) and the chain rule for partial differentiation. Which is more complex but still manageable with applied thought and practice.

Mike