Expert Systems:-

7.1. Given a first order logic knowledge base constructed as follows:-

TELL(KB, AT(1,1))

TELL(KB, AT(2,1) => DEAD( ))

TELL(KB, AT(3,1) AND NOT DEAD( ) => GOAL( ))

TELL(KB, X,
Y, AT(X,Y) AND MOVE(A,B)

AND ( A = 1 OR A = -1 OR A = 0)

AND ( B = 1 OR B = -1 OR B = 0) => AT(X+A, Y+B) )

What is the solution to the following query ?

ASK(KB, GOAL( ) )

Show your proof.

7.2. Given a situation calculus knowledge base as shown:-

TELL(KB, AT(1,1, S0))

TELL(KB, AT(2,1, Sn) => DEAD(Sn ))

TELL(KB, DEAD(Sn) => DEAD(Sn+1))

TELL(KB, AT(3,1,Sn) AND NOT DEAD( Sn) => GOAL(Sn ))

TELL(KB, X,
Y, AT(X,Y,Sn) AND MOVE(A,B,Sn)

AND ( A = 1 OR A = -1 OR A = 0)

AND ( B = 1 OR B = -1 OR B = 0) => AT(X+A, Y+B,Sn+1) )

What is the solution to the following query ?

ASK(KB, S, GOAL(S ) )

Show your proof.

Planning:-

8.1. Given a situation calculus knowledge base as shown:-

TELL(KB, AT(1,1, S0))

TELL(KB, AT(1,2, Sn) => DEAD(Sn ))

TELL(KB, AT(3,1, Sn) => DEAD(Sn ))

TELL(KB, AT(2,3, Sn) => DEAD(Sn ))

TELL(KB, DEAD(Sn) => DEAD(Sn+1))

TELL(KB, AT(1,3,Sn) AND NOT DEAD( Sn) => GOAL(Sn ))

TELL(KB, X,
Y, AT(X,Y,Sn) AND MOVE(A,B,Sn)

AND{ ( A = 1 AND B = 1) OR (A = -1 AND B = 0)

OR (A = -1 AND B = 0) OR (A = 0 AND B = 1)}

=> AT(X+A, Y+B,Sn+1) )

Is there a solution to the following query ?

ASK(KB, p, GOAL(PlanResult(p, S0 ) )

Show your proof.

8. 2. What is the difference between the hierarchical decomposition (HD) planner and the HD partial order planner (POP) ?

8. 3. Explain why the downward and upward solution property must
hold
for the HD-POP algorithm to be efficiently implemented.

Machine Learning:-

Past Exam Questions

Natural Language:-

10.1. Briefly define the terms *speaker*, *listener* and
*utterance*
in terms of natural language communication.

10.2. What are the categories of utterance that agents may generate ?

10.3. Discuss the ways that communication can be implemented between two agents.

10.4. Discuss the problems associated with general natural language
identifying the reasons why it is a research problem that still has to
be solved.

Machine Vision & Pattern Recognition:-

11.1. In terms of pattern recognition define the terms *feature
extraction*
and *classification*.

11.2. Define the terms *supervised training* and *unsupervised
training*.

11.3. What is *overtraining* and *undertraining* and how
can
it be avoided.

11.4. What is the problem with using Bayesian Theory in pattern classification ?

11.5. What are the main features of the nearest neighbour classification algorithm ? Discuss the efficiency and optimality of this algorithm.

11.6. How does the K - nearest neighbour algorithm improve on the
nearest
neighbour classification algorithm ?