Philosophy, Computing, and AI

Each assignment is worth 10 points. Several have multiple parts. Answer all parts. Quote each part of the prompt in bold. Write your answer below it.

So, for example, in Assignment #1, the first part is

• Explain the difference between knowledge and belief. Which is more fundamental for understanding the intelligence of a rational agent. State and explain your reasons for your view.

What you submit for this part of the assignment should take the following form:

Explain the difference between knowledge and belief.

.... [your answer] ....

Which (knowledge or belief) is more fundamental for understanding the intelligence of a rational agent.

.... [your answer] ....

You should write your answers so that a reader who does not already know the material can understand what you have written. The best way to do this is to provide detailed answers.

Submissions that do not follow these instructions will not receive credit.

★ Assignment #1 (four parts)

• Explain the difference between knowledge and belief. Which is more fundamental for understanding the intelligence of a rational agent. State and explain your reasons for your view.

• It seems possible that the intelligence of a rational agent can be understood computationally and that the reasoning in which the agent engages can be understood as a form of computation. Explain how deliberation about whether a given end is achievable might be understood as a form of computation. Be sure to give a detailed example to illustrate your explanation.

• Consider the following logic program

          a ← b.
          b ← c.
          c.

Suppose the query to this program is

          ?-a.

In posing this query, the question is whether a is a logical consequence of the program.

Set out the steps in the backward chaining computation that begins when the query is put to the program. (The steps in the backward chaining computation are the matching procedure described in the lectures. So your answer should set out the steps that occur in this computation.) If you can, explain how the steps in this computation correspond to a reductio ad absurdum proof in the propositional calculus.

• Suppose that I form the belief that some object I see is red because the object looks red to me. It is possible that "the object looks red to me" is true and that "the object is red" is false. (The object might be white and have a red light illuminating on it.) Does this possibility show that that my belief that the object is red is not rational? Be sure to explain your answer in detail.

★ Assignment #2 (three parts)

• In Prolog, load a logic program that consists in the fact that

man(socrates).
(In English: Socrates is a man.)

and the rule that

mortal(X):-man(X).
(In English: All men are mortal.)

With respect to this program, ask whether something is a man

?- man(X).

and ask whether something is mortal

?- mortal(X).

Submit a screen shot or shots to prove that you have posed these queries to the program.

• With respect to the logic program in the prior question, set out the steps in the backward chaining computation (the matching procedure with unification) to answer the second query.

• Complete one of the following:

Present and explain the Wason Selection Task and the observed experimental results. Subjects do poorly in this experiment but better in formally equivalent tasks with more meaningful content. Does this show that it is a mistake to use logic programming to model the intelligence of a rational agent? Give reasons to support your view.

or

Present and explain the Byrne Suppression Task and the observed experimental results. Do the observed results in this experiment show that it is a mistake to use logic programming to model the intelligence of a rational agent? Again, be sure to give reasons to support your view.

★ Assignment #3 (two parts)

• Use the "Fox and Crow" example to explain how the agent gets an achievement goal in the logic programming/agent model. Present the example and give the explanation in detail.

• Explain why achievement goals are generally not logical consequences of the agent's KB.

★ Assignment #4 (four parts)

• Explain the difference between the classical negation introduction rule and the negation-as-failure rule. In your explanation, be sure to provide a clear statement of the two rules. Be sure to explain what defeasible reasoning is, and be sure to explain which rule is a rule of defeasible reasoning.

• Consider the following simple logic program

          a ← b, c.
          b ← d.
          b ← e.
          c.
          d ← h.
          e.
          f ← g, b.
          g ← c, k.
          j ← a, b.       

Suppose the query is

          ?- f. 

Set out the steps in the computation (the matching procedure) to answer this query.

• Suppose that the above program allows negation-as-failure and that in the program

         g ← c, k.  

is replaced with

         g ← c, not-k. 

Suppose the query to this new program is

          ?- f.  

Set out the steps in the computation (the matching procedure) to answer this query. Be sure to explain the relevant differences between this computation and the prior one.

• Explain Kowalski's attempt to explain the experimental results in the Suppression Task in terms of the logic programming/agent model. Be sure to give a detailed explanation. To do that, you must present the relevant episodes of reasoning in the Suppression Task as computations involving logic programs. Finally, explain whether you think Kowalski's response to the Suppression Task is adequate. Give reasons for your view.

★ Assignment #5 (two parts)

• In the logic programming/agent model, prohibitions function to rule out certain plans to bring about a goal the agent seeks to achieve. Use the "Runaway Trolley" example to explain how prohibitions work in the logical programming/agent model. Be sure to give a detailed explanation.

• In the logic programming/agent model, integrity constraints function to rule out certain explanations. Use the following "medical" example to explain what abductive reasoning is and how integrity constraints work in the context of the logic programming/agent model.

The beliefs and integrity constraint in the "medical" example are

          bronchitis ← influenza.
          bronchitis ← smokes.
          coughing ← bronchitis.
          wheezing ← bronchitis.
          fever ← influenza.
          soreThroat ← influenza.

          false ← smokes ∧ nonsmoker.

For abduction, the possible hypotheses are

          smokes, influenza. 

Suppose that the agent observes

          wheezing.  

What is the explanation or explanations for this observation? Set out the steps in the computation for determining this explanation or explanations.

Suppose instead that the agent observes

          wheezing, nonsmoker.  

Now what is the explanation? Explain the role the integrity constraint plays in determining this explanation. Set out the steps in the computation that involves the integrity constraint.

★ Assignment #6 (three parts)

• Consider the language defined by following grammar and lexicon:

          s -> np vp 
          np -> det n 
          vp -> v np 
          vp -> v 
          det -> a 
          det -> the 
          n -> woman 
          n -> man 
          v -> loves

Is the string of words 'the woman loves a woman' a sentence of the language? If it is a sentence of the language, provide a parse tree for the string.

• Put the following rules and facts into a prolog program:

          s(X,Z) :- np(X,Y), vp(Y,Z). 
          np(X,Z) :- det(X,Y), n(Y,Z). 
          vp(X,Z) :- v(X,Y), np(Y,Z). 
          vp(X,Z) :- v(X,Z). 

          det([the|W],W). 
          det([a|W],W). 

          n([woman|W],W). 
          n([man|W],W). 

          v([loves|W],W).     

Provide a screen shot to show that you have done it. Pose the following query

          ?- s([a,good,man,is,hard,to,find],[]).  

Provide a screen shot to show that you have done it. If you can, set out the steps involved in answering this query including the unifications that occur.

• Consider the following rule from the lexicon discussed in the lecture notes:

          common_noun(park,X) :- park(X).  

What does this rule say? Be sure to give your answer in terms of words, grammatical categories, and extensions. Consider the query

          ?- common_noun(park,tree03). 

What question is this query asking? Again, be sure to give your answer in terms of words, grammatical categories, and extensions.

★ Assignment #7 (one part)

• Explain the most important points in Kowalski's attempt to explain the experimental results in the Selection Task in terms of the logic programming/agent model. Explain whether you think Kowalski's response is adequate. Be sure to give your explanations and reasons in detail.