Home

Homework 6

Table of Contents

Task 1 (20 pts)

Assume we have the following Prolog database:

parent(tom, liz).
parent(tim, bob).
parent(bob, ann).
parent(bob, pat).
parent(pat, jim).

male(tim).
male(tom).
male(bob).
male(jim).
female(pam).
female(liz).
female(pat).
female(ann).

mother(X) :- parent(X, _), female(X).
  1. Write Prolog queries for the following questions:
    1. Who is Ann's parent?
    2. Is Jim Bob's child?
    3. Who is Jim's grandparent?
  2. Write a rule that describes a child, so that child(X, Y) means X is the child of Y. Don't allow a person to be his or her own child.
  3. Define an ancestor rule. An ancestor is a parent of a person, a grandparent, a great-grandparent, etc. Naturally, this rule is recursive (like the member predicate in the notes).
  4. Define a "has no parents" rule. Use negation ("negation as failure").

Make sure you actually test your queries and rules with SWI-Prolog! And look at the video from Day 13, about the 14 minute mark, for help.

Task 2 (10 pts)

Do the following unify, and if so, what are values for the variables that make the unification work? Note: Each pair of terms has new variable assignments. Look at the video from Day 14 for help.

  1. parent(tom, liz) and parent(tom, X)
  2. parent(X, liz) and parent(liz, X)
  3. parent(X, Y) and parent(Y, X)
  4. head(X, [a, b, c]) and head(b, [a, b, c])
  5. head(foo(Y), X) and head(foo(Z), [foo(w)])
  6. foo(bar(a, b), baz(c)) and foo(bar(X), baz(Y))

Task 3 (10 pts)

Draw a resolution tree for mother(X), refering back to the database above. Note that there is a particular order by which Prolog creates subgoals (that is, the order that facts/rules are written in the knowledge base). You only need to show the tree until the first time the predicate is proved. Look at the video from Day 14 for help.

Task 4 (60 pts)

Extend the Prolog program at the bottom of this page to support the following interactions. (The user types the stuff after the |:) Look at the video from Day 15, starting at about the 18 minute mark, for help. In particular, you'll need to use assertz (21:30 mark in the video). Write this in a Prolog source file (hw6.pl maybe), and test your code. Submit the source file so the grader can also test your code.

?- consult('nldb.pl').
%    library(error) compiled into error 0.00 sec, 18,352 bytes
%   library(lists) compiled into lists 0.01 sec, 46,096 bytes
%  library(readln) compiled into readln 0.01 sec, 60,696 bytes
% nldb.pl compiled 0.01 sec, 78,560 bytes
true.

?- run.
|: quit.
Ok we're done! 
true.

?- run.
|: anteater is bigger than ant.
Ok, I learned that ant is smaller than anteater 
|: ant is smaller than foot.
Ok, I learned that ant is smaller than foot 
|: what is bigger than ant?
These are bigger: 
anteater, foot.
|: what is smaller than anteater?
These are smaller: 
ant.
|: what is smaller than ant?
These are smaller: 

|: godzilla is bigger than anteater.
Ok, I learned that anteater is smaller than godzilla 
|: what is bigger than ant?
These are bigger: 
anteater, foot, godzilla.
|: what is smaller than godzilla?
These are smaller: 
anteater, ant.
|: today is thursday.
Ok, I learned that today is a/an thursday 
|: socrates is mortal.
Ok, I learned that socrates is a/an mortal 
|: what is socrates?
socrates is a/an: 
mortal.
|: what is today?
today is a/an: 
thursday.
|: what is xyz?
I'm not sure! Care to tell me? 
|: is xyz abc?
No 
|: xyz is abc.
Ok, I learned that xyz is a/an abc 
|: is xyz abc?
Yes
|: what is 15?
15 
|: what is 2 + 2?
4 
|: what is 2 - 3 * 5?
-13 
|: what is 2 * 3 - 2?
2 
|: what is 4 / 3?
1.3333333333333333 
|: wow, you are very smart!
Huh? 
|: quit.
Ok we're done! 
true.

?-

Start with this code. Pay special attention to the "TODO" comments.

:- use_module(library(readln)).

% these allow us to use assertz with these two predicates
:- dynamic(learned_isa/2).
:- dynamic(learned_smaller/2).

% strip_puncation/2 & read_sentence/1 from The Art of Prolog 2nd Ed. (Sterling & Shaprio)

strip_punctuation([], []).
strip_punctuation([Word|Tail], [Word|Result]) :-
    \+(member(Word, ['.', ',', '?', '!'])),
    strip_punctuation(Tail, Result).
strip_punctuation([_|Tail], Result) :-
    strip_punctuation(Tail, Result).

read_sentence(Input) :-
    readln(Input1),
    strip_punctuation(Input1, Input).

reply([Head|Tail]) :- write(Head), write(' '), reply(Tail).
reply([]) :- nl.

% TODO: code reply_list/1

run :- read_sentence(Input), run(Input), !.

run(['quit']) :- reply(['Ok we\'re done!']).

run(Input) :-
    process_sentence(Input),
    read_sentence(NextInput),
    !, run(NextInput).

process_sentence([what, is, bigger, than, Object]) :-
    % TODO: code findall using Object and creating Things
    reply(['These are bigger:']),
    reply_list(Things).

process_sentence([what, is, smaller, than, Object]) :-
    % TODO: finish predicate

process_sentence([what, is | Math]) :-
    math(Math, Z),
    reply([Z]).

process_sentence([what, is, Object]) :-
    % TODO: finish predicate

process_sentence([what, is, _]) :-
    reply(['I''m not sure! Care to tell me?']).

process_sentence([is, Subject, Object]) :-
    % TODO: put a line of code here
    reply(['Yes']).

% TODO: handle case where "is xyz abc" says "No"
% (i.e., above predicate failed)

process_sentence([Subject, is, Object]) :-
    assertz(learned_isa(Subject, Object)),
    reply(['Ok, I learned that', Subject, 'is a/an', Object]).

% TODO: handle "x is bigger than y" and "x is smaller than y"
% similarly to how "is" is handled above,
% using assertz(learned_smaller(...))

process_sentence(_) :-
    reply(['Huh?']).

% right-associative math

% default math case (an atom)
math(X, Z) :-
    number(X),
    Z = X.

% default math case (a list)
math([X], Z) :-
    number(X),
    Z = X.

math([X, + | Tail], Z) :-
    % TODO: finish predicate

% TODO: code rest of math

isa(X, Y) :- learned_isa(X, Y).

% TODO: code transitive isa

smaller(X, Y) :- learned_smaller(X, Y).

% TODO: code transitive smaller
Intro to AI material by Joshua Eckroth is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License. Source code for this website available at GitHub.