COGS 100 Lecture Notes - Lecture 10: Marvin Minsky, John Searle, Turing Test
Document Summary
Intelligent agents intelligent software capable of mimicking human intelligent behavior: intelligent systems consist of multiple intelligent agents, autonomic self managing, approaches in cogsci. Our brains compare perceptual inputs to expectations: according to him, the primary function of intelligence is prediction. If you"re watching a movie, both branches work at the same time from the bottom. Parallel processing: auto-associative recall for object recognition, case, difficult to express in terms of rules, physical symbol system ex: rule based systems and other models. Therefore, perception is very difficult: systems, gps, the rules consist of if-then assertion. Neural network approach: warren mcculloch assembled research group with walter pitts and jerry. Activation given value at a time: 2) In absence of activation, a preciously activated node decays back to inactive state with a resting state: 3) Example iam model: t": activation has time and value given by ai(t, in absence of activation, previously activated nodes decay back to inactive resting state.