Computers and Reality - Cognitive Science 110 - Spring 2007

1. Explain what excitatory and inhibitory connections are.
2. Define what preferred stimulus means.
3. On page 39, line 16, it says: "... or of something similar to it." What do you think "similar" means here? If the pattern looks exactly the same except that it's shifted down one row, or rotated 90 or 180 degrees, would it be considered "similar" in the context of this discussion?
4. On page 40, line -3 (i.e., third line from the bottom), it says: "... higher resolution, both in space and in brightness, ...". What does this mean?
5. On page 42, line 9: what does "two-layer" mean? How many neurons are there in each layer?
6. Churchland says there is something that is biologically realistic about this ANN and how it learns. What is it? (See page 42).
7. Page 42, line -11, "... serial computer...". What does "serial" mean?
8. Explain in detail how the "error" in the network's output is computed (page 43). Note: a "vector" is just a list of numbers; e.g., [5, 8, 0, -3] is a vector.
9. Page 43: explain how squaring the differences "exaggerates the relative importance of larger errors over smaller ones."
10. Page 44, line 3: what does "connective configuration" mean?
11. Page 46, line 1: "... relatively large but far from exhaustive...". Explain "far from exhaustive". Why does Churchland think it's "far"?
12. Page 47, line 15: "... we can reconstruct ...". Explain how.