what is artificial intelligence

25
What is artificial intelligence? shreya chakraborty

Upload: shreya-chakraborty

Post on 06-Jul-2015

285 views

Category:

Engineering


3 download

DESCRIPTION

introduction to artificial intelligence for engineering

TRANSCRIPT

Page 1: What is artificial intelligence

What is artificial intelligence?

shreya chakraborty

Page 2: What is artificial intelligence

The physical symbol system

•set of entities==> symbols

•All symbols/instances related in some physical way.

•Processes(creation, modification, reproduction and destruction)

shreya chakraborty

Page 3: What is artificial intelligence

Intelligence requires knowledgeKnowledge-

1. Voluminous

2. Hard to categorize

3. Constantly changing

4. Organisation different to usage

shreya chakraborty

Page 4: What is artificial intelligence

What is an AI technique?

•Should capture generalisation

•Understood by people who provide it

•Easily modified to correct errors and reflect changes

•Mostly accurate

•Overcome bulk possibilities to produce result

shreya chakraborty

Page 5: What is artificial intelligence

3 important AI techniques?Search

Use of knowledge

Abstraction

shreya chakraborty

Page 6: What is artificial intelligence

4 steps to solve a problem•Define problem precisely

•Analyse problem

•Isolate and represent task knowledge necessary

•Choose best problem solving technique

shreya chakraborty

Page 7: What is artificial intelligence

State space representation

• Basis of AI methods

• Structure:-

• Formal definition for problem

• Explore space trying to find path from current state to goal state

shreya chakraborty

Page 8: What is artificial intelligence

State space problem

•Define state space (all possible configurations of relevant objects)

•Specify initial state

•Specify goal state

•Specify set of rules that define actions

shreya chakraborty

Page 9: What is artificial intelligence

Production System•Rules : Applicability ->Operation

•Knowledge/Databases

•Control strategy

•Rule applier

shreya chakraborty

Page 10: What is artificial intelligence

Control Strategy requirements•It should cause motion

•It should be systematicshreya chakraborty

Page 11: What is artificial intelligence

Breadth-first

Searchshreya chakraborty

Page 12: What is artificial intelligence

Depth-first

Searchshreya chakraborty

Page 13: What is artificial intelligence

Heuristic SearchHeuristic Knowledge incorporated in search

• in rules themselves

• or as a function

shreya chakraborty

Page 14: What is artificial intelligence

Heuristic function• Problem description

measures(numbers)

• F(x)=g(x)+h’(x)

shreya chakraborty

Page 15: What is artificial intelligence

Problem characteristics•Decomposable?•Solution steps can be undone?•Problem’s universe predictable?•Good solution obvious?•Desired solution a state or a path?•Large amt of knowledge absolutely required to solve the problem/•Can computer take problem and return solution?

shreya chakraborty

Page 16: What is artificial intelligence

Issues in Search Program Design

*instead of building entire tree, programs represent trees in rules implicitly, and generate what needs to be explored

*forward vs backward reasoning

*rule matching

*knowledge representation problem and frame problem

shreya chakraborty

Page 17: What is artificial intelligence

Heuristic Search techniques1. Depth First

2. Breadth Firstshreya chakraborty

Page 18: What is artificial intelligence

3. Generate and Test

• Generate solution

• Check to see if actually a solution by comparison

• If solution found quit, else repeat all steps

shreya chakraborty

Page 19: What is artificial intelligence

4. Hill Climbing

• Simple Hill Climbing

• Steepest Ascent Hill Climbing

shreya chakraborty

Page 20: What is artificial intelligence

5. Best First Search

• Or-graphs

• A* algorithm

shreya chakraborty

Page 21: What is artificial intelligence

using REDUCE-AND

· Use REDUCE on each immediate subgoal until there are

no more subgoals, or until REDUCE finds a subgoal that

is not satisfied.

· If REDUCE has found a subgoal that is not satisfied,

announce that the goal is not satisfied; otherwise,

announce that the goal is satisfied.

using REDUCE-OR

· Use REDUCE on each subgoal until REDUCE finds a

subgoal that is satisfied.

· If REDUCE has found a subgoal that is satisfied,

announce that the goal is satisfied; otherwise, announce

that the goal is not satisfied.

shreya chakraborty

Page 22: What is artificial intelligence

6. Problem Reduction

• And-or graphs

• AO* algorithm

shreya chakraborty

Page 23: What is artificial intelligence

8. Constraint Satisfaction

shreya chakraborty

Page 24: What is artificial intelligence

8. Constraint Satisfaction

shreya chakraborty

Page 25: What is artificial intelligence

9.Mean Ends Analysis

To perform means-ends analysis,· Until the goal is reached or no more procedures areavailable,- Describe the current state, the goal state, and

the difference between the two.- Use the difference between the current state

and goal state, possibly with the description ofthe current state or goal state, to select apromising procedure.

- Use the promising procedure and update thecurrent state.

· If the goal is reached, announce success; otherwise,announce failure.

shreya chakraborty