Artificial Neural Network | Types | Feed Forward | Feedback | Structure | Perceptron | Machine Learning | Applications

What is Neural Network?

Neural network is a computer system modelled on the human brain and nervous system. Artificial Neural Network (ANN) is an information processing paradigm based on the working of biological nervous systems to process information. It is similar to human brain. It is composed of a large number of highly interconnected processing elements called neurons. Neurons work in unity to solve specific problem. Like human being,  ANN also learn by example. We can configure an Artificial Neural Network for a specific application such as pattern recognition or data classification through a learning process. Learning in ANN is similar to biological systems. It also include adjustments to the synaptic connections that exist between the neurons.

Biological vs artificial neuron MSA-Technosoft

Artificial Neural Network (ANN) is based on a collection of interconnected nodes called artificial neurons. It is a simplified version of biological neurons in an animal brain. Each connection acts as a synapse between artificial neurons. It is used to transmit signals from one to another. The artificial neuron that receives the signal first process it and then pass the signal to artificial neurons connected to it.

Artificial Neural Network is developed with the belief that working of human brain can be imitated by making the right connections. Silicon and wires can be used as living neurons and dendrites.

Structure of an Artificial Neural Network

The human brain is made up of 86 billion nerve cells. These nerve cells are called neurons. These neurons are connected to other thousand cells by Axons. Input or Stimuli from external environment via sensor organs are accepted by dendrites. These inputs create electric impulses. These impulses quickly travel through the neural network. A neuron then send the message to other neuron to handle the issue or does not send it forward.

Artificial Neural Network (ANN) are composed of multiple nodes. These nodes act as biological neurons of human brain. The neurons are linked together and they interact with each other. The nodes can take input data and perform simple operations on the basis of that data. The result of these operations is passed to other neurons. The output at each node is called its activation or node value. Each link contain its own weight. ANNs are capable of learning, which takes place by altering weight values.

Perceptron

Perceptron is a type of artificial neuron. The first artificial neuron was produced in 1943 by the neurophysiologist Warren McCulloch and the logician Walter Pits. Perceptrons were developed in the 1950s and 1960s by a scientist Frank Rosenblatt. He was inspired by earlier work of Warren McCulloch and Walter Pitts.

A perceptron takes several binary inputs, x1, x2 ,…, and produces a single binary output as shown in diagram below:

http://neuralnetworksanddeeplearning.com/images/tikz0.png

Here, the perceptron has three inputs, x1, x2, and x3. Rosenblatt proposed a simple rule to compute the output. He introduced weights, w1, w2, …These are real numbers expressing the importance of the respective inputs to the output. The neuron gives output in terms of 0 or 1. It is determined by whether the weighted sum ∑jwjxj is less than or greater than some threshold value. Threshold is also a real number. It is a parameter of the neuron. We can represent it in algebraic form as follows:

Output = {0,1if ∑jwjxj≤ thresholdif ∑jwjxj> threshold(1)

Why use neural networks?

Neural network has ability to derive meaning from complicated or imprecise data. It can be used to extract patterns and detect trends that are too complex to understand by human or other computer techniques. A well trained neural network work as an expert for the 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.

Neural network has Adaptive learning capacity. It is ability to learn how to do tasks based on the data given for training or initial experience.

An Artificial Neural Network can create its own organisation or representation of the information it receives during learning time.

An Artificial Neural Network provide real-time operations. Its computations may be carried out in parallel. Special hardware devices are being designed and manufactured to take advantage of this capability.

It has the capacity of 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.

Types of Neural Networks

Feed Forward Neural Network

Signals travel in one way i.e. from input to output only in Feed forward Neural Network. There is no feedback or loops. The output of any layer does not affect that same layer in such networks. Feed forward neural networks are straight forward networks that associate inputs with outputs. They have fixed inputs and outputs. They are mostly used in pattern generation, pattern recognition and classification.

feed forward neural network MSA-Technosoft

Feedback Neural Network

Signals can travel in both the directions in Feedback neural networks. Feedback neural networks are very powerful and can get very complicated. Feedback neural networks are dynamic. The ‘state’ in such network keep changing until they reach an equilibrium point. They remain at the equilibrium point until the input changes and a new equilibrium needs to be found. Feedback neural network architecture is also referred to as interactive or recurrent, although the latter term is often used to denote feedback connections in single-layer organisations. Feedback loops are allowed in such networks. They are used in content addressable memories.

feedback neural network MSA-Technosoft

Neural Network Machine Learning

Machine learning is the branch of computer science.it is done with the help of data and algorithm. Machine learning algorithms use training sets of real-world data instead of relying on human instructions to infer models that are more accurate and sophisticated than humans could devise on their own. Neural networks act as a subset of algorithms built around a model of artificial neurons spread across three or more layers here. Deep learning is generally used to describe particularly complex networks with many more layers than normal in neural networks. The networks are able to develop much greater levels of abstraction, which is necessary for some complex tasks like image recognition and automatic translation.

An Artificial Neural Network is capable of learning. It need to be trained. There are several learning strategies as follows:

Supervised Learning:

In this type of machine learning the training dataset contains inputs data and the value you want to predict.

The ANN will use the training data to learn a link between the input and the outputs. The idea behind is that the training data can be generalized and that the ANN can be used on new data with some accuracy.

It need a teacher that is scholar than the ANN itself. We can take an example of a teacher and student here. The teacher feeds some example data about which the teacher already knows the answers.

Pattern recognition is one of the best example of such learning. The ANN comes up with guesses while recognizing. Then the teacher provides the ANN with the answers. The network then compares it guesses with the teacher’s “correct” answers and makes adjustments according to errors.

Unsupervised Learning:

This type of learning is required when there is no example data set with known answers. For example, searching for a hidden pattern. In this case, clustering i.e. dividing a set of elements into groups according to some unknown pattern is carried out based on the existing data sets present.

Reinforcement Learning:

In this type of learning the best ways to earn the greatest reward is to find. Rewards can be winning a game, earning more money or beating other opponents. They present state-of-art results on very human task. This strategy built on observation. The ANN makes a decision by observing its environment. If the observation is negative, the network adjusts its weights to be able to make a different required decision the next time.

Neural Network Applications

Artificial Neural Network (ANN) is based on the processing of human brain. It is developed to simplify tasks that are easy for human but difficult for machines. The algorithms can be used to model complex patterns and prediction problems with the help of ANN.

  • Telecommunications: Image and data compression, automated information services, real-time spoken language translation
  • Electronics: Code sequence prediction, IC chip layout, chip failure analysis, machine vision, voice synthesis.
  • Anomaly Detection: can be trained to generate an output when something unusual occurs that misfits the pattern.
  • Speech: Speech recognition, speech classification, text to speech conversion.
  • Signal Processing: Neural networks can be trained to process an audio signal and filter it appropriately in the hearing aids.
  • Medical: Cancer cell analysis, EEG and ECG analysis, prosthetic design, transplant time optimizer
  • Aerospace: Autopilot aircrafts, aircraft fault detection.
  • Industrial: Manufacturing process control, product design and analysis, quality inspection systems, welding quality analysis, paper quality prediction, chemical product design analysis, dynamic modelling of chemical process systems, machine maintenance analysis, project bidding, planning, and management.
  • Software: Pattern Recognition in facial recognition, optical character recognition, etc.
  • Automotive: Automobile guidance systems.
  • Time Series Prediction: ANNs are used to make predictions on stocks and natural calamities.
  • Signal Processing: Neural networks can be trained to process an audio signal and filter it appropriately in the hearing aids.
  • Military: Weapon orientation and steering, target tracking, object discrimination, facial recognition, signal/image identification.
  • Financial: Real estate appraisal, loan advisor, mortgage screening, corporate bond rating, portfolio trading program, corporate financial analysis, currency value prediction, document readers, credit application evaluators.
  • Transportation: Truck Brake system diagnosis, vehicle scheduling, routing systems.
  • Control: ANNs are often used to make steering decisions of physical vehicles.

Must share your views on the comment section.  For more informative technology related blogs, must go through Tech-Blogs of MSA Technosoft here https://msatechnosoft.in/blog/category/tech-blogs/

Leave a Reply

Your email address will not be published. Required fields are marked *

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