ilp-exercises.tex

\setlength{\medskipamount}{1.6\medskipamount}%%%%% Mona's suggestion

%%%% 19.2: Knowledge in Learning (1 exercises, 1 labelled)
%%%% =====================================================

\begin{exercise}[dbsig-exercise]%
Show, by translating into conjunctive normal form and applying
resolution, that the conclusion drawn on \pgref{dbsig-page} concerning
Brazilians is sound.
\end{exercise} 
% id=19.0 section=19.2


%%%% 19.4: Learning Using Relevance Information (2 exercises, 0 labelled)
%%%% ====================================================================

\begin{uexercise}
For each of the following determinations\index{determination}, write down the logical
representation and explain why the determination is true (if it is):
\begin{enumerate}
\item Design and denomination determine the mass of a coin.
\item For a given program, input determines output.
\item Climate, food intake, exercise, and metabolism determine weight gain
and loss.
\item Baldness is determined by the baldness (or lack thereof) of
one's maternal grandfather.
\end{enumerate}
\end{uexercise} 
% id=19.1 section=19.4

\begin{iexercise}
For each of the following determinations\index{determination}, write down the logical
representation and explain why the determination is true (if it is):
\begin{enumerate}
\item Zip code determines the state (U.S.).
\item Design and denomination determine the mass of a coin.
\item Climate, food intake, exercise, and metabolism determine weight gain
and loss.
\item Baldness is determined by the baldness (or lack thereof) of
one's maternal grandfather.
\end{enumerate}
\end{iexercise} 
% id=19.1 section=19.4

\begin{exercise}
Would a probabilistic version of determinations be useful? Suggest a
definition.
\end{exercise} 
% id=19.2 section=19.4


%%%% 19.5: Inductive Logic Programming (4 exercises, 3 labelled)
%%%% ===========================================================

\begin{exercise}[ir-step-exercise]%
Fill in the missing values for the clauses \(C_1\) or \(C_2\) (or both) in the
following sets of clauses, given that \(C\) is the resolvent of \(C_1\)
and \(C_2\):
\begin{enumerate}
\item \(C = \J{True} \implies P(A,B)\), \(C_1 = P(x,y) \implies Q(x,y)\), \(C_2
= ??\).
\item \(C = \J{True} \implies P(A,B)\), \(C_1 = ??\), \(C_2 = ??\).
\item \(C = P(x,y) \implies P(x,f(y))\), \(C_1 = ??\), \(C_2 = ??\).
\end{enumerate}
If there is more than one possible solution, provide one example of
each different kind.
\end{exercise} 
% id=19.3 section=19.5

\begin{exercise}[prolog-ir-exercise]%
\prgex Suppose one writes a logic program that carries out a
resolution\index{resolution} inference step. That is, let \(\J{Resolve}(c_1,c_2,c)\) succeed
if \(c\) is the result of resolving \(c_1\) and \(c_2\). Normally, \(\J{Resolve}\)
would be used as part of a theorem prover by calling it with \(c_1\) and
\(c_2\) instantiated to particular clauses, thereby generating the
resolvent \(c\). Now suppose instead that we call it with \(c\)
instantiated and \(c_1\) and \(c_2\) uninstantiated. Will this succeed in
generating the appropriate results of an inverse resolution step?
Would you need any special modifications to the logic programming
system for this to work?
\end{exercise} 
% id=19.4 section=19.5

\begin{exercise}[foil-literals-exercise]%
Suppose that \prog{Foil} is considering adding a literal to a clause
using a binary predicate \(P\) and that previous literals (including
the head of the clause) contain five different variables. 
\begin{enumerate}
\item How many
functionally different literals can be generated? Two
literals are functionally identical if they differ only in the names
of the {\em new} variables that they contain.
\item Can you find a
general formula for the number of different literals with a predicate
of arity \(r\) when there are \(n\) variables previously used?
\item Why does \prog{Foil} not allow literals that contain no
previously used variables?
\end{enumerate}
\end{exercise} 
% id=19.5 section=19.5

\begin{uexercise}
Using the data from the family tree in \figref{family2-figure}, or a
subset thereof, apply the \prog{Foil} algorithm to learn a definition
for the \(\J{Ancestor}\) predicate.
\end{uexercise} 
% id=19.6 section=19.5




\resetmedskipamount