Neural Networks

Neural Networks

1. A computing system made up of a number of simple, highly interconnected processing elements, which process information by their dynamic state response to external inputs

2. Composed of a large number of highly interconnected processing elements (neurons)

3. Advantages

   • Adaptive learning
   • Self-Organization
   • Real Time Operation
   • Fault Tolerance via Redundant Information Coding

4. Similar to Brain

5. Units

   • Nodes
   • Links
   • Activation
   • Weight
   • Activation Function'
   • Bias Weight

6. Structures

   • Feed-forward networks
     • Signals travels one way
   • Feedback Networks
     • Signals travels both ways via loop

7. Learning

   • Supervised
     • Presented inputs together with outputs and tries to produce the result as close as the provided output
   • Unsupervised
     • Network response to the input patterns and adjusts its weight accordingly