SCXT 350

 

Final Exam

 

 

 

 

 

Name ___________________________________

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Thursday, May 13

12:00 PM – 2:00 PM

200 pts.


 

1.         (10 pts.)           We began the semester with an assertion of Descartes that we would never be fooled by an automaton that appeared as a human.  What was “Descartes Test” and why did he believe that no automaton would ever pass it?

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.         (20 pts.)           Turing, in his 1950 paper, gave us a response to Descartes in his description of the “Turing Test” and his belief that it would eventually be passed. 

 

            a)         Describe the Turing Test (making sure not to describe instead the “Imitation Game”).

 


            (continuation of problem 2)

 

            b)         If a computer program passed the Turing Test, how would this be a response to Descartes?

 

 

 

 

 

 

 

 

 

 

 

 

 

3.         (5 pts.) In our early discussions of knowledge representation systems, we discussed the idea of a grammar.  How does a grammar give a response to Descartes?

 

 

 

 

 

 


 

4.)        (15 pts.)           Searle has a response to Turing (with a response to Shank’s Yale group as well).  Describe Searle’s Chinese Room experiment (including what the inputs and outputs are), and why Searle thinks of this as an argument against the Turing Test.

 


5)         (20 pts.)           One of the tenets of cognitive science is that we can view thinking in terms of information processing.  David Marr gave us a way to view information processing on several levels.  Describe the three levels, and also say something of how one level constrains the other two.  Illustrate your discussion by using a typical hand calculator as an example.


6)         (10 pts.)           In the following list of disciplines that make up cognitive science, say at which of Marr’s three levels each provides an explanation, together with a brief (one sentence) explanation as to why you think so.  There could potentially be several reasonable answers – pick one that you like.  The justification is important.

 

                        philosophy

 

 

 

 

                        psychology

 

 

 

 

                        computer science generally

 

 

 

 

                        symbolic artificial intelligence

 

 

 

 

 

                        artificial neural networks

 


 

A chief tenet of cognitive science is that cognition is computational.  The next several questions explore this idea:

 

7)         (10 pts.)           What is an algorithm?

 

 

 

 

 

 

 

 

8)         (5 pts.)             What is the relationship (if any) between a correctly written computer program and an algorithm?

 

 

 

 

 

 

 

9)         (5 pts.)             Informally, what does it mean to say that something is Turing-computable?

 

 

 

 

 

 

 

10)       (5 pts.)             What is the Turning-Church hypothesis?

 

 

 

 

 

11)       (5 pts.)             Combining your answers to problems 9 – 10 above, what does this say about the statement “cognition is computable”?

 


12)       (10 pts.)           While we think of computers as very powerful machines, there are really only about five basic tasks that a computer can perform.  Please list them below.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

One way to test the idea that cognition is computable is to try and write programs which perform tasks that we think of as demonstrating intelligence.  A difficult job!  The area of computer science which studies this is called Artificial Intelligence, which is in turn divided into two camps:

 

13)       (10 pts.)           What is a connectionist architecture?  That is, what approach do connectionists take in trying to write programs which appear to exhibit intelligent behavior?


14)       (10 pts.)           Draw a picture of a two-input (x,y)  perceptron which fires when x = 0 and y = 1, and fails to fire in all other cases.  Carefully note the weights on each of the inputs (including the bias element).  Although it might be a good idea to start with a picture of a line, what I will be looking for is the diagram (picture) of the perceptron.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


 

15)       (15 pts.)           The other approach to AI is called symbolic AI, or GOFAI.  The commitment of symbolic AI is to the physical symbol system hypothesis.  Briefly describe a physical symbol system and state the physical symbol system hypothesis.

 


16)       (15 pts.)           One sometimes writes that AI = KR + Search.  In thinking of problem solving as search, we think of states, operators, preconditions, and postconditions.  Give brief definitions of these terms, using the blocksworld microworld as a source of examples for your definitions.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 


 

 

17)       (5 pts.)             A student looking for the room in which an exam is to be given runs from door to door looking for the room.  In confronting a dead end, the student backs up to the last point at which there was an alternate route to take.  What sort of search is the student using?

 

 

 

 

 

 

 

 

 

 

 

 

18)       (5 pts.)             A kind soul tells the student of the previous problem that the exam is being given on one of the upper levels of the building.  What has the kind soul just handed the student in aid of the student’s search?

 

 

 

 

 

 

 


 

19)       (10 pts.)           We have examined several knowledge representation (KR) schemes, most recently a slot-and-filler structure in CLIPS.  What is a slot-and-filler structure?  As a part of your explanation, give an example in CLIPS.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20)       (10 pts.)           Some knowledge is stored in productions.  In CLIPS, write a production (rule) which says that if we have in our database a fact saying

 

                        (animal (motion approaching) (teeth big) (food-status hungry) (size big))

 

            then we should add to our database the fact

 

                        (run away)