CS607 - Artificial Intelligence
Course Page
Q & A
Short Question & Answers
Q1: Write advantages of Artificial Neural Networks.
Excellent for pattern recognition, excellent classifiers, handles noisy data well, well for generalization
Q2: Write disadvantages of Artificial Neural Networks.
Parallel architecture: The power of Artificial Neural Networks is lie in their parallel architecture, most machines are serial (Von Neumann architecture)
Defined rules: Lack of defined rules to build a neural network for a specific problem as there are too many variables, for instance, the learning algorithm, number of neurons per layer, number of layers, data representation
Knowledge is implicit:
Data dependency:
Q3: What is machine learning?
Machine learning is the ability of knowledge based systems to improve through experience and is divided into three categories: Rote learning, Inductive learning and deductive learning.
Q4: What intelligence ability is better in living species than in computers? Why it is better?
Computers are better than people are at remembering things exactly and at performing complex numeric calculations with speed. But human brains still beat computers in a number of ways. For one, humans can integrate information from many different variables and stimuli, and they can learn by experience, observation and experimentation. Computers can't easily adapt to changing situations. Sure, they can be programmed to perform outstandingly in a particular field, but they are not able to function in multiple disciplines. Moreover, the things that make humans truly unique (emotion, empathy, self-awareness, ambition) are beyond the capacity of computers.
Q5: Discuss ID3 in decision tree representation.
ID stands for interactive dichotomize. The first step of ID3 is to find the root node. It uses a special function GAIN, to evaluate the gain information of each attribute. For example if there are 3 instances, it will calculate the gain information for each. Whichever attribute has the maximum gain information, becomes the root node. The rest of the attributes then fight for the next slots.
Q6: Define ESDLC stages.
Feasibility study, Rapid prototyping, Alpha system (in-house verification), Beta system (tested by users), Maintenance and evolution
Q7: Discuss POP algorithm.
1. The partial-order planning algorithm – POP:
2. POP(initial_state, goal, actions) returns plan
3. Begin
4. Initialize plan ‘p’ with initial_state linked to goal state with two special actions, start
   and finish
5. Loop until there is not unsatisfied pre-condition
6. Find an action ‘a’ which satisfies an unachieved pre-condition of some action ‘b’ in
    the plan
7. Insert ‘a’ in plan linked with ‘b’
8. Reorder actions to resolve any threats
9. End

In POP algorithm pre-conditions of finish action are not met. We just backtrack by adding actions that meet these unsatisfied pre-condition predicates. New unsatisfied preconditions will be generated for each newly added action. Then we try to satisfy those by using appropriate actions. we keep on doing that until there is no unsatisfied precondition.
Q8: What are the linear model’s phases? Describe any 5
The main phases of the linear sequence are Planning, Knowledge acquisition and analysis, Knowledge design, Code, Knowledge verification, System evaluation
Q9: What is robotics? Write the name of different active areas involved in robotics?
Robotics is the highly advanced and totally hyped field of today. Literally speaking, robotics is the study of robots. Robots are complex combination of hardware and intelligence, or mechanics and brains.

Active areas involved in robotics: Robotics is truly a multi-disciplinary area, having active contributions from, physics, mechanics, biology, mathematics, computer science, statistics, control theory, philosophy, etc.
Q10: Differentiate find s and candidate elimination algorithms.
Find-S: Find-S is guaranteed to output the most specific hypothesis h that best fits positive training examples. The hypothesis h returned by Find-S will also fit negative examples as long as training examples are correct.

Candidate-Elimination: Candidate-Elimination gives Outputs a description of set of all hypotheses consistent with the training examples.
Q11: How to apply implication method.
In implication method we use the degree of support for the entire rule to shape the output fuzzy set. The consequent of a fuzzy rule assigns an entire fuzzy set to the output. This fuzzy set is represented by a membership function that is chosen to indicate the qualities of the consequent. If the antecedent is only partially true, i.e., is assigned a value less than 1, then the output fuzzy set is truncated according to the implication method. One rule is usually not enough so 2-3 rules implemented by it. The output of each rule is a fuzzy set. The output fuzzy sets for each rule are then aggregated into a single output fuzzy set. Finally the resulting set is defuzzified, or resolved to a single number.
Q12: How working memory is related to the Expert System?
Working memory contains the problem facts that are discovered during the session’ according to Durkin. One session in the working memory corresponds to one consultation.
Q13: What is the basic task of clustering?
The basic task of clustering is to identify and group similar individual data elements based on some measure of similarity. So basically using clustering algorithms, classification information can be produced from a training data.
Course Instructor

Dr. Zafar Alvi PHD (EE)
University of Bradford

Artificial Intelligence: Structures & Strategies for Complex Problem Solving by George F Luger

Artificial Intelligence by Elaine Rich, Kevin Knight

Artificial Intelligence by Patrick Henry Winston