Turing on Computing Machinery and Intelligence
1. Is anything wrong with the question "Can machines think?"
2. Is winning the "imitation game" (passing the Turing test) a necessary condition for computer intelligence? Is it a sufficient condition?
3. Which objection to artificial intelligence discussed by Turing do you find most plausible? Can you find ways of overcoming the objection?
4. Are computers now any closer to being conscious now than they were in Turing's day? Does it matter?
5. Can computers learn?
6. What obstacles to computer intelligence did Turing neglect?
Newell and Simon on symbolic intelligence
1. What is the physical symbol hypothesis?
2. What is the evidence for it?
3. Are there plausible alternatives to this hypothesis?
4. Is problem solving accomplished by heuristic search? Are there alternative conceptions of problem solving?
5. Does animal intelligence depend on symbol systems and heuristic search?
6. Do symbols in computers have any meaning? Does it matter whether they relate to the world?
7. How do brains produce symbols?
Minsky on Frames
1. What is a frame?
2. How do frames provide understanding?
3. How do frames produce learning?
4. How much of our knowledge do you think can be represented in frames?
5. What do you think of Minsky's criticism of the logistic approach?
6. What are the strengths and weaknesses of Minsky's frame theory as an account of the structure and growth of human knowledge?
Dreyfus and Searle vs. AI
1. Why does Dreyfus doubt that Winograd's program SHRDLU understands language?
2. Why is Dreyfus skeptical of Minsky's frames? Are his criticisms effective?
3. Dreyfus' says his basic point is that since intelligence must be situated it cannot be separated from the rest of human life. What does this mean? What are its consequences?
4. Do thought experiments like Searle's Chinese room ever demonstrate anything?
5. What do you think is the most plausible reply to Searle? Is it effective?
6. Why does Searle think it is crucial for symbols to have intentionality?
7. What causal powers must a brain or machine have in order for its symbols to possess semantics as well as syntax?
Rumelhart and Smolensky on neural networks
1. According to Rumelhart, what is brain-style computation? How does it differ from traditional computation?
2. What are constraint satisfaction problems? Why is connectionism a natural approach to solving them?
3. How can neural networks learn? How much of human learning can connectionism explain?
4. According to Smolensky, what are the differences between the symbolic and subsymbolic paradigms?
5. What is subsymbolic computation?
6. How are schemata implemented in neural networks?
7. What seem to you to be the strengths and weakness of connectionism as an approach to understanding human thinking?
Fodor & Pylyshyn and Ramsay et al. on connectionism
1. What are the most appropriate levels of explanation for human cognition?
2. Why do Fodor and Pylyshyn think that connectionism is incompatible with classical approaches to cognition? Are they right?
3. What do F&P think are the most serious shortcomings of connectionism? Do you agree?
4. Is connectionism at best a theory of implementation?
5. What is eliminativism, according to Ramsay et al.?
6. Does connectionism imply eliminativism?
7. Is common sense psychology based on propositional attitudes worth
saving?
Brooks on intelligence without representation
1. Why does Brooks think that AI researchers are guilty of self-deception?
2. How do Brooks' Creatures differ from previous AI programs?
3. Does Brooks really want to do away with representation?
4. Why should a network of simple machines be intelligent?
5. Why is embodiment critical for intelligent systems?
6. Does Brooks' research program answer the challenges to AI of Dreyfus and Searle?
Van Gelder on the Mind as a Dynamic System
1. How does the Watt governor differ from a computational governor?
2. Is the Watt governor a plausible analog for mental machinery?
3. How are dynamical systems different from computational systems?
4. Are connectionist systems computational, dynamical, or both?
5. Does the dynamical systems view of mind constitute a plausible alternative to the computational view?
Chalmers on Consciousness
1. What is the hard problem of understanding consciousness? What is its relation to the easy problem?
2. Is research in cognitive psychology and neuroscience relevant to solving the hard problem?
3. What "extra ingredients" might explain consciousness?
4. What is the difference between reductive and non-reductive explanations of consciousness?
5. Does Chalmers have a promising theory of consciosness?
6. Could robots be conscious?
Week 11
Schank & Cleary on creativity
1. What do Schank and Cleary think are the key processes in creativity?
2. Would these processes explain creativity in science? In art? In literature?
3. What are explanation patterns? How useful are they?
4. What are the toughest problems remaining for the psychological and computational study of creativity?
5. Why are some people much more creative than others?
6. Can computers only do what people tell them to do?
7. Can computers be as creative as people?
Week 12
Velásquez on emotion
1. Why does Velásquez want a computational model of emotions?
2. Is the Cathexis cognitive model psychologically plausible?
3. How are emotion and behavior systems related?
4. Does Yuppy really learn emotions?
5. Could robots have emotions rather than merely simulating emotions? Do human emotions require a human body?
6. What role do emotions play in decision making? What role should they play?
7. A human is a:
(a) rational animal.
(b) emotional animal.
(c) both of the above.
(d) neither of the above.