Q1. Write answers to following in brief. (any 13)                                    (13)

1.What is the relation between percept sequence and action in intelligent agents?

2.What is the need for performance measures in intelligent agents?

3.How would one measure the rationality of an agent?

4.Give a single example to explain beta cutoff used in game playing.

5.Why iterative deepening is considered better then minimax for real world problems?

6.How could Samuel’s checkers playing program beat himself after some time?

7.What was the problem with single layer perceptron? Why they failed?

8.What is the difference between square and sigmoid activation function?

9.How can we see that not all the inputs are gained by one single unit in unsupervised learning?

10.What is tangled hierarchy? How can inferential distance solve the problem?

11.How is class different from other sets?

12.How many type of attribute the class can have? Give example of each of the type of attributes.

13.Explain intersection search with example.

14.What is the meaning of membership grade?

15.Why when MB is positive MD must be zero and vice versa?

16.Who is knowledge engineer?

17.What is inference engine?

18.What is expert behavior?

19.Why explanation facility is an important component of an Expert system?

Q2. Solve any two. (12)

1.Describe a frame system for GLS. Include any five classes and three instances of your liking and try to provide following.

a.Use of Multiple Inheritance

b.Use of is-covered-by and mutually-disjoint with

c.Pick up one or more slots and explain following terms with example.

i.Domain

ii.Range

iii.Transfers through

iv.To compute

v.Single valued

2.Draw a semantic nets for following statements

a. “Atal’s chair is snached by manmohan”

b.All Indians are proud of Anil’s achievement

c.Everybody believes in some form of God.

3.How can we solve a problem of signature recognition using backpropagation algorithm and multilayer perceptron? Do not write back propagation algorithm, write steps involved in solving the problem including how input and output units will be chosen, how would each input be provided, how training takes palce and how the solution is obtained and how testing inputs are provided etc.

4.How would following statements represented in NMRS? How will IN and OUT list be maintained? How would CP justification be used to identify contradiction?

a.I am in India (default)

b.I am in hilly area (observation)

c.I am on the bank of river flowing westwards. (observation)

d.The river is Bramhputra. (a and c)

e.I am in Assam (a and d)

f.I am to travel in west to reach Delhi (e)

g.Predominant language in Assam is Assmese (fact)

h.The signboards at the bank of the river are in Urdu (observation)

5.Draw an agent for chess playing program. How can you make your agent as following?  

a.Simple reflex agent

b.Model based reflex agent

c.Goal based agent

d.Utility based agent

e.Learning agent