They sometimes claim an intelligent behavior of an artificial neural network. As far as I know, however, they do not usually look so intelligent -- it might be successful but always in a same deterministic way. Human intelligence is more spontaneous. Sometimes it tries to make a task in yet another way, just for a change, even if it's not a most efficient way. Currently I'm interested in how we can implement this flexibility to an artificial neural network. Spiking Timing Dependency Plasticity (STDP) of spiking neurons might be one such idea.






"Finding a Needle in a Haystack: From Baldwin Effect to Quantum Computation." A. Imada (2007) Keynote Speech. Proceedings of the International Conference on Computer Information Systems and Industrial Management Applications. pp. (28 June - 30 June, 2007. University of Finance and Management, Elk, Poland) 

"Hinton & Nowlan's computational Baldwin effect revisit: Are we happy with it?" A. Imada (2007), Proceedings of the International Conference (Informacijos Mokslai 42-43 2007) associated with Kompiuterininku Dienos 2007 (KoDi'07) pp. 207 -- 212. (13 - 15 September 2007. KTU Panevezio Institutas, Panevezys, Lithania) 

"Can a Learning Robot Survive in a Desert? -- A 2-D expansion of the Jeep Problem." A. Imada (2007), 14th International Multi-conference: Advanced Computer Systems (ACS-2007) and Artificial Intelligence, Software Technologies Biometrics and Information Technology Security (AISBIS-2007) (17-19 October, 2007. Miedzyzdroje, Poland)