Slide 14 of 37
Notes:
Perhaps parochial to view chess playing and similar activities as the most important aspects of human achievement.
Probably incorrect to assume that solution of problems like chess playing will automatically lead to solution of such problems as speech recognition, spatial navigation, sensory-motor coordination and balance, recognition of objects, path planning, obstacle avoidance, learning from experience, selective attention, anticipation, and effective response to changing environmental conditions.
There is no reason to believe that modeling a fly landing on a ceiling should be intrinsically more difficult than modeling the thought processes of a scientist.
Differences in our ability to address these two problems are more likely to be due to differences in effort allocated and in the mechanisms exploited for their solution than to differences in the difficulty of the problems.
“The way it looked at the time [1970] was that the perceptron people had t do an immense amount of mathematical analysis and calculating to solve even the most simple problems of pattern recognition, such as discriminating horizontal from vertical lines in various parts of the receptive field, while the symbol-manipulating approach had relatively effortlessly solved hard problems in cognition, such as proving theorems in logic and solving combinatorial puzzles” (Dreyfus & Dreyfus, 1988, p. 24).
