Introduction to Artificial Neural Network
An Artificial Neural Network (ANN) is an information processing system that is inspired by how the biological nervous systems work, such as the brain, process information. Like human, Artificial Neural Network learn from example to solve a problem without being programmed with any task specific rules.
Use of Artificial Neural Networks
Artificial Neural Networks can derive meaning from complicated or imprecise data, which can be used to extract patterns and detect trends that are too complex to notice by either humans or other computer techniques. A trained Artificial Neural Network can be thought as an “expert” in that category of information it has been given to analyse. This expert can then be used to provide projections given new situations of interest and answer “what if” questions.
Some of other advantages of Artificial Neural Network includes –
- Adaptive learning: An ability to learn from example how to do tasks based on the data given for training or initial experience.
- Self-Organization: An Artificial Neural Network can create its own organization or representation of the information it receives during learning time.
- Real Time Operation: Artificial Neural Network computations may be carried out in parallel, and special hardware devices are being designed and manufactured which take advantage of this capability.
- Fault Tolerance via Redundant Information Coding: Partial destruction of a network leads to the corresponding degradation of performance. However, some network capabilities may be retained even with major network damage.
With these advantages, Artificial Neural Network have some disadvantages also. These are discussed at below.
Basic Structure of an Artificial Neural Network
The idea of ANNs is based on the belief that working of human brain by making the right connections, can be imitated using silicon and wires as living neurons and dendrites. It is composed of large number of highly interconnected processing components that works together to solve a problem. The components are called as Node in Artificial Neural Network and these node imitate the biological neurons of human brain into a computer. As the neurons of human brain interconnected through links to interact with each other, in ANN nodes are also interconnected with each other to process data. These node can take input data & process those data to solve a simple problem and passed the result to an another node. Then with the help of this output another node solve another problem and again produce a result. These way the all the nodes together solve a complex problem. The output of each node is called as activation or node value. In Artificial Neural Network each link is associated with a weight. ANNs are capable of learning, which takes place by altering weight values. The following illustration shows a simple Artificial Neural Network.
Types of Artificial Neural Networks
There are mainly six types of Artificial Neural Network topologies. These are discussed below.
- Feedforward Neural Network: This is the simplest among all the Neural Network Topologies. Here, the information is move in one direction only. The information is pass from the input nodes to output nodes directly or through hidden nodes if any. It has no cycle or loop in it.
- Radial Basis Function Network: It is an artificial neural network that uses radial basis functions as activation functions. The output of thIS network is a linear combination of radial basis functions of the inputs and neuron parameters. Radial basis function networks have many uses including function approximation, time series prediction, classification, and system control etc.
- Recurrent Neural Network: Unlike the Feedback Neural Network, the Recurrent Neural Network is a neural network where data can be flow in any direction i.e. bidirectional. This allows it to exhibit dynamic temporal behavior for a time sequence. A Recurrent Neural Network has the ability to use it’s internal memory to process arbitrary sequence of input. This neural network is a popular choice for tasks such as handwriting and speech recognition.
- Modular Neural Network: Human brain operates as a large collection of small networks. Like this, in Modular Neural Network, several small network corporate with each other to solve a big problem. Each small individual network take input and produce an output. Then intermediary network accepts the inputs of each of these individual neural networks, processes them, and creates the final output for the modular neural network. Here, the independent neural networks do not interact with each other directly.
- Physical Neural Network: A physical neural network is a type of artificial Neural Network in which an electrically adjustable resistance material is used to emulate the function of a neural synapse. “Physical” neural network is used to emphasize the reliance on physical hardware used to emulate neurons as opposed to software-based approaches which simulate neural networks. While the physical hardware emulates the neurons, the software emulates the neural network.
- Kohonen Self-Organizing Neural Network: It is Invented by Teuvo Kohonen. The Kohonen self-organizing neural network is ideal for the visualization of low-dimensional views of high-dimensional data. The self-organizing neural network is different from other neural networks and applies competitive learning to a set of input data, as opposed to error-correction learning applied by other neural networks. The Kohonen self-organizing neural network is known for performing functions on unlabeled data to describe hidden structures in it.
Disadvantages of Artificial Neural Network
- The neural network requires training to operate it.
- Requires high processing time for large neural networks.
- The architecture of a artificial neural network is different from the architecture and history of microprocessors so they have to be emulated.