Where are the limits ?
Alicia HERAZ
I want to ask an important question about the role of a computer science student doing researchs in artificial intelligence.
I'm in artificial intelligence lab, we make research to improve intelligent tutoring systems. Knowing that cognitive sciences are a subdomain of artificial intelligence, and that cognitive sciences needs computer science, bilology, psychology and neurosciences. How can I improve some thing in cognitive sciences ? as Computer science student and not as psychologue or biologist.
Somtimes, there are some questions which are not yet resolved in psychology, like the exact definition of emotions, or the exact theory of emotional induction. If there is several theories, what and how to choose one of them and why ? Psychologist tests to measure some difficult aspects, are there reliable ? Can we use a psychologist test to validate a system ?
http://alicia.heraz.free.fr/
Étudiante en Doctorat aux Laboratoires HERON et GRITI
(514) 343 6532 + 45042
Nouveau : téléphonez moins cher avec Yahoo! Messenger ! Découvez les tarifs exceptionnels pour appeler la France et l'international. Téléchargez la version beta.
John Flynn
Well, it seems one way to measure and evaluate progress on a computer simulation of human behavior would be to use yourself as the subject. If you try to simulate yourself, even if only in a limited sense, you can then test the resulting computer simulation by developing a relevant set of stimuli. You then document your personal reaction to each of the stimuli. Next you provide the same set of stimuli to the simulation and document its reaction to each. You can then compare your reactions to the reactions from the computer simulation. If they are the same you are probably on the right track and can then add additional complexity to the simulation and repete the testing/evaluation process by appropriately increasing the scope of the stimuli.
Brandon Franklin
-Brandon
--
My LinkedIn Profile: http://www.linkedin.com/profile?viewProfile=&key=1615193
Harris
After working in Pattern REcognition for a decade or so, I believe that
strong AI is not yet something we can comprehend fully and implement it
as a mathematical model or computer simulation. Instead, we know how to
deal with very difficult problems at a lower level, for example
extracting and understanding a #D model from stereoscopic image
acqusition, or "simulating" the way we make decisions from uncertain
inputs or contradicting data.
If you are interested in Cognitive Sciences more than dealing with
"everyday" AI problems (example: medical diagnostics and how a doctor
thinks), perhaps Computer Science is helpful for you only at the very
"high" part of AI or simply as a tool for statistical analysis and data
mining. If you are interested more in Computer Science and how you can
"embed" cognitive aspects into it, I would suggest you take a look into
areas like Pattern Recognition (ANN, SVM, etc) and standard AI (state
machines, automata, fuzzy logic, etc).
Good luck :-)
--
Harris
www.di.uoa.gr