Quick Links
| University Papers | University Syllabus | Entrance Exam |
| PSU Papers | Bank Papers | Placement Papers |
| VTU | Anna Univerity Syllabus | Anna Univerity Papers |
Anna University Question Paper
Computer Science and Engineering
CS2351-Artificial Intelligence
Sixth Semester
May/June 2014 Question Paper
ANNA UNIVERSITY SYLLABUS: CLICK HERE
Part A(10*2=20)
1.Give the structure of an agent n an environment.
2.List the criteria to measure the performance of search strategies.
3.P=>Q=>-PVQ Construct a truth table to show that this equivalence holds.
4.Write the generalized modulus ponens rule.
5.List out the various planning techniques.
6.What is contingency planning?
7.List down two applications of temporal probabilistic models.
8.Define unertainity.
9.List some applications where reinforcement learning is used.
10.What are the three factors involved in the analysis of efficiency gains from EBL(Explanation Based Learning).
Part B(5*16=80)
11.a)1)What are the problems caused due to incomplete knowledge on the states or
actions?Define each with example. (8)
2)Decsribe constraint satisfaction problem in detail. (8)
Or
b)1)Explain the components of problem definition with an example. (8)
2)Briefly explain the search strategies in uniformed search. (8)
12.a)1)What are the steps to convert First order logic sentence to Normal form?Explain ach step. (6)
2)Represent the following sentences in predicate logic and convert the following sentences to CNF form.
1)All women who like ice-creams like chocolates. (2)
2)No man is happy with a spendthrift wife. (2)
3)The best movie in Hollywood is always better than the movie in Bollywood. (2)
4)Some people like eating outside all the time and Some people like eating at home all
the time. (2)
5)It might be argued that one aspect of intelligent behavior is the ability to infer new facts
about the world by combining existing ones .Has the theory of logic given us a tool to allow computers to display the sort of intelligence? Can Humans make other laps of inference that are impossible with logic alone? (2)
Or
b)1)Differentiate propositional logic with FOL .List the inference rules along with suitable examples for First order Logic. (10)
2)Consider the following sentences.
1)John likes all kind of foods.
2)Apples are food.
3)Chicken is food.
4)Anything anyone eats and isn’t killed alive.
5)Sue eats everything bill eats.
A)Translate these sentences into formulas in predicate logic.
B)Convert the formulas of a part into clause form.
C)Prove that “John likes peanuts”using forward chaining.
D)Prove that “John likes peanuts”using backward chaining
13.a)1)Give a detailed account on the planning with state-space research. (8)
2)Explain the concept behind partial order planning with examples. (8)
Or
b)1)Explain the process of modifying the planner for decompositions with suitable examples. (8)
2)Dscribe planning graph in detail. (8)
14)a)1)Define uncertain knowledge, prior probability and conditional probability .State the Baye’s theorem.How it is useful for decision making under uncertainty about knowledge? Explain the method of performing exact inference in Bayesian networks
briefly. (6)
2)What is Bayesian networks? How is the Bayesian networks used in representing uncertainty about knowledge? Explain the method of performing exact inference in
Bayesian networks. (10)
Or
b)1)Describe the role of Hidden Markov Model in speech recognition. (8)
2)Consider the following facts. (8)
1) I saw my cat in the living room in 3 hours ago.
2) 2 hours ago my door below open.
3)Three quarters of the time my door belows open,my cat runs outside the door.
4) One hour ago I thought I heared a cat noise in my living.
Assume I was half cetain.
5)In one hour period the probability that the cat will leave the room is 0.2 .There is
also a 0.2 probability that he may enter the room. What is the certainity that the cat is my
living room? Use Bayesian networks to answer this.
15)a)1)Suppose you set this as a machine learning problem to a learning agent. You specify that the positives are the exams which are difficult. (10)
1)Which would the most specific hypothesis that the agent would initially
construct?
2)How would the agent generalize the hypothesis in light of the second positive
example? What are the other possible generalizations are there?
3)How would you use the 2 hypothesis from parts (1) and (2) to predict whether
an exam will be difficult?
4)What score would the 2 hypothesis get in terms of predictive accuracy over the
training set,and which one would be chosen as the learned hypothesis? (6)
Or
b)1)Consider a simple domain : waiting at a traffic light.Give an example of a decision
tree for this domain (8)
1)Make a list of relevant variables.
2)Explain how we can use the concept of information or expected information
gain to determine which ariale to choose,for a maximally Compact decision tree.
2)For the case of learning to play tennis( or some other sport which you are familier). Is
this supervised learning or reinforcement learning?
0 Comments