My Guide to Intelligent Systems Viva

Chapter 1: Types of Agents

We outline four basic kinds of agent program that embody the principles underlying almost all intelligent systems:

1. Simple Reflex Agent

2. Model Based Agent

3. Goal-Based Agent

4. Utility-Based Agent

We will try to understand each of these taking the help of some real world and well acquainted set of examples.

Eg 1) Simple Reflex Examiner:

Examiner: What is the algorithm used for finding solution to 8-puzzle problem?

You: I think...

Examiner: Wrong!

Simple Reflex Agent gives a predefined action for a given percept.

Eg 2) Model Based Examiner:

Examiner: What is the algorithm used for finding solution to 8-puzzle problem?

You: I think we can use the hill-climbing algorithm.

Examiner: How?

You: First we find the heuristic value for initial stage...

Examiner: Stop! I knew you would start off babbling whatever you have mugged up when I ask this question. Tell me any other real world example for hill-climbing algorithm.

You: (stare into space)

Examiner: I knew you students do not bother to learn but only see through classes notes for viva.................

Model Based Agent knows how the world evolves and reacts to the agent's actions.

Eg 3) Goal Based Examiner:

Examiner: What is the algorithm used for finding solution to 8-puzzle problem?

You: I think we can use the hill-climbing algorithm.

Examiner: How?

You: First we find the heuristic value for initial stage...

Examiner: Stop! I knew you would start off babbling whatever you have mugged up when I ask this question. Tell me any other real world example for hill-climbing algorithm.

You: (stare into space)

Examiner: I knew you students do not bother to learn but only see through classes notes for viva. Tell me, can you apply the A* algorithm to find optimal path to the hill-climbing algorithm tree?

You: The A* algorithm is always optimal as it uses the cost plus heuristic value........

Examiner: See! You only know the theory. This proves that you are not interested in learning the concepts at all!

Goal Based Agent knows how the world evolves and reacts to the agent's actions and also those actions lead to a certain Goal.

Eg 4) Utility Based Examiner:

Examiner: What is the algorithm used for finding solution to 8-puzzle problem?

You: I think we can use the hill-climbing algorithm.

Examiner: How?

You: First we find the heuristic value for initial stage...

Examiner: Stop! I knew you would start off babbling whatever you have mugged up when I ask this question. Tell me any other real world example for hill-climbing algorithm.

You: (stare into space)

Examiner: I knew you students do not bother to learn but only see through classes notes for viva. Tell me, can you apply the A* algorithm to find optimal path to the hill-climbing algorithm tree?

You: Examiner, the A* algorithm is always optimal as it uses the cost plus heuristic value........

Examiner: I am not at all happy with your answers! On a scale of 1-10 the wrongness of your answers is 12!

Utility Based Agent maps the correctness of the result to a real value.