Artificial Intelligence - Download as PowerPoint

Document Sample
Artificial Intelligence - Download as PowerPoint Powered By Docstoc
					Artificial Intelligence
               Definition

 The science of making computers do
  things that would require intelligence if
  done by people.
 This definition is given by Marvin
  Minsky, who is considered the father of
  AI
                Brief History
 1950’s : The term ‘Artificial Intelligence’ was
  coined at The Dartmouth Conference.
 1970’s : Expert Systems -- Useful Offshoot
 1980’s : Biologically Inspired Models
 1990’s : Biologically Inspired Models continue.
     First Applications of AI

 Game Playing
 No longer considered intelligent activity
 Merely a fast searching mechanism
 Searching for the best alternative as fast
  as possible.
              Branches of AI
   Robotics
   Machine Vision
   Natural Language Interpretation
   Voice Recognition
   Game Playing
   Pattern Recognition
   Machine Learning
   Neural Networks
   Genetic Algorithms
   Expert Systems
             Roots of AI

 Computer Science
 Cognitive Psychology
 Philosophy
 Electrical Engineering
 Linguistics
 Management Science
 Biology
 Neural Sciences
    AI Applications in Business
   Credit Analysis
   Bankruptcy Prediction
   Bond Rating
   Auditor Judgment
   Commodity Trading
   Asset Allocation
   Corporate Merger Prediction
   Time Series Forecasting
   Stock Price Prediction
   Risk Analysis
         Computers Vs. Humans
      Computers                      Human beings

   Process data sequentially      Process data in parallel
   Processing very fast           Processing speed slow
   Very good in arithmetic        Poor and slow in arithmetic
   Cannot process spatial         Can process spatial data
    data well                       well
   Extremely accurate             Not very accurate
   Cannot see, touch, feel        Can see, touch, feel
   Cannot recognize               Can recognize patterns,
    patterns, handwriting,          handwriting, continuous
    continuous speech               speech
     Artificial Neural Networks

 Artificial Neural Networks is a biologically
  inspired technique to solve many types of
  real world problems.
 ANN mimic the architecture of the human
  brain and simulate the processing of the
  brain on a sequential processing
  computer.
     The Architecture of ANN

                    Output Layer
    Processing
    Units
                       Hidden Layer

Weights
                             Input Layer
And on and on

				
DOCUMENT INFO