skip to Main Content

Artificial Neural Network Basic Concepts

Artificial Neural Network Basic Concepts

Artificial Neural Network Basic ConceptsDefinition of Artificial Neural Network Basic Concepts – An artificial neuron network (ANN) is a computing System model based on the structure and functions of biological neural networks. The system learn to do tasks without specific programming. The ANN is a collection or group of interconnected units called Artificial Neurons. The neurons are organised in layers. Each interconnection between neurons transmit a unidirectional signal with strength. If the combined incoming signals are strong, the receiving end neuron activate and propagate signal to the downstream neurons interconnected to it. The main moto of the neural network approach is to solve problems as human brain can do.

The inventor of the first neurocomputer was Dr. Robert Hecht-Nielsen, according to him “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”.

Types of Artificial Neural Networks (ANN)

There are two types of ANN model – Feed Forward and Feedback.

Feedforward ANN –

In this model, the information is flowing in unidirectional mode. The communication between transmitter and receiver is not maintained due to unidirectional propagation of information. These type of model is used in Pattern Generation, Recognition and Classification. There is no feedback loops exists in this type of model. This type of model have fixed inputs and outputs.

Feedback ANN – The looping is allowed in this model. The model also contain feedback loop which is helpful during content addressable memories use.

Application Areas of Artificial Neural Networks (ANN)

  1. Medical Areas− Cancer cell analysis, EEG and ECG analysis, transplant time optimizer and others.
  2. Electronics and Telecommunications areas− real-time spoken language translation, Image and data compression, automated information services, Code sequence prediction, IC chip layout, chip failure analysis, machine vision, voice synthesis.
  3. Speech recognition, speech classification, text to speech conversion.
  4. Time Series Prediction − ANNs are used to make predictions on stocks and natural calamities.
  5. Signal Processing − Neural networks can be trained to process an audio signal and filter it in the hearing aids.
  6. Software − Facial Pattern recognition, optical character recognition
  7. Financial aids – Exist as loan advisor, mortgage screening, corporate bond rating, portfolio trading program, corporate financial analysis, currency value prediction, document readers, credit application evaluators.

Limitation of Artificial Neural Network (ANN)

The Backpropagation Neural Network properties act like a black boxes. In fact the user has no other choice than to feed it input and await for output. The Network is also slower than other types of network which kill time of CPU.

Characteristics of Artificial Neural Networks (ANN)

  1. Choice of Model – simpler to complex.
  2. Robustness
  3. Learning & Simplified algorithms

This is all about Artificial Neural Network Basic Concepts, in next article we will discuss regarding another Network applications.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back To Top