Creative work undertaken on a systematic manner to increase the stock of knowledge, intelligence of human, society and culture that used to create to new application using Machine Learning. The main purpose of scientific research is discovery, documentation, interpretation and innovation research of methods and systems for advancement of human knowledge.Artificial Neural Network is a computational model used in machine learning or scientific research which is based on large collection of simple units Called artificial neurons. The Artificial neural networks are used to solve the wide variety of tasks, such as Speech recognition, Image processing, Computer vision, bio metrics, Prediction system, Recommendation system, data mining and deep learning.

Recurrent Neural Network (RNN)

A recurrent neural network is a network in which neurons sends feedback signals to one another.A RNN has loops in them that allow information to be carried across neurons while reading in input. These links allow the activations from the neurons in a hidden layer to feed back into themselves at the next step in the sequence.

In other words, at every step, a hidden layer receives both activation from the layer below it and also from the previous step in the sequence. This structure essentially gives recurrent neural networks memory.

Convolutional Neural Networks (CNN)

Convolutional Neural Networks are biologically-inspired variants of MLPs.It is a type of feed-forward artificial neural network in which the connectivity pattern between its neurons is inspired by the organization of the animal visual cortex.It is comprised of one or more convolutional layers and then followed by one or more fully connected layers as in a standard multilayer neural network.

Architecture of a CNN is designed to take advantage of the 2D structure of an input image. Main benefit of CNNs is that they are easier to train and have many fewer parameters than fully connected networks with the same number of hidden units.

Feedforward Neural Networks

It is the simplest type of the artificial neural network in which information moves only in one direction. Here information goes from input nodes to the hidden nodes and then from it to the output nodes. Therefore, there are no loops exist in this type of network.

Basically it is a biologically inspired classification algorithm. Every unit in a layer is connected with all the units in the previous layer. During normal operation, that is when it acts as a classifier, there is no feedback between layers. This is why they are called feedforward neural networks.

Stochastic Neural Networks

Stochastic neural networks are a kind of artificial neural networks built by presenting random variations into the network, either by giving the network’s neurons stochastic transfer functions, or by giving them stochastic weights.

It is very helpful to resolve optimization problems, since the random fluctuations help it escape from local minima.Boltzman learning is the best example of this type of network. Each neuron is binary valued, and the chance of it firing depends on the other neurons in the network. Stochastic neural networks have initiate applications in oncology,risk management, bioinformatics, and other similar fields.

Modular Neural Networks

It is a model when Different models of neural networks combined into a single system to form modular neural networks. Each single network is made into a module that can be freely intermixed with modules of other types in that system.Modular neural network architecture builds a bigger network by using modules as building blocks.

The architecture of a single module is simpler and the subnetworks are smaller than a monolithic network. Due to the structural modifications the task the module has to learn is in general easier than the whole task of the network. The modules are independent to a certain level which allows the system to work in parallel.

Simple Recurrent Networks

The simple recurrent network is a selected version of the Backpropagation neural network that makes it feasible to manner of sequential input and output. It is normally a three-layer network where a duplicate of the hidden layer activations is stored and used as enter to the hidden layer inside the subsequent time step. The preceding hidden layer is completely related to the hidden layer. Because the community has no recurrent connections thee whole network can be trained with the backpropagation algorithm as normal.

It can be trained to read a series of inputs into a output pattern, to generate a series of outputs from a given input pattern, or to map an input series to an output sequence.

Physical Neural Networks

A physical neural network is a category of synthetic neural network wherein an electrically adjustable resistance material is used to emulate the characteristic of a neural synapse. "Physical" neural network is used to highlight the reliance on physical hardware used to emulate neurons rather than software program-based totally techniques which simulate neural networks.

A physical neural network as one or more nonlinear neuron-like nodes used to sum signals and Nano connections formed from nanowires nanoparticles, or nanotubes which regulate the motion strength input to the nodes. It has very significant application in nanotechnology field.

Spiking Neural Networks

The field of artificial spiking neural networks is an attempt to emphasize the neurobiological aspects of artificial neural computation.As an artificial neuron models the relationship between the inputs and the output of a neuron, artificial spiking neurons describe the input in terms of single spikes, and how such input leads to the generation of output spikes.

The transmission of a single spike from one neuron to another is mediated by synapses at the point where the two neurons interact. An input or presynaptic spike arrives at the synapse, which in turn releases neurotransmitter which then influences the state.

Dynamic Neural Networks

In dynamic networks, the output depends not only on the current input to the network, but also on the current or previous inputs, outputs, or states of the network.

The dynamic neural structures, in general, can be classified into two categories:

The first category encompasses the dynamic neural structures developed based on the concept of single neuron dynamics as an extension of static neural networks.

The second category encompasses dynamicneural structures which are developed based on the interaction of excitatory and inhibitory or antagonistic neural subpopulations.Dynamic neural networks not only deal with nonlinear multivariate behaviour, but also include (learning of) time-dependent behaviour such as various transient phenomena and delay effects.

Cascading Neural Networks

It is a type of the supervised learning. Cascade-forward networks are similar to feed-forward networks, but include a connection from the input and every previous layer to following layers. During the training process, neurons are selected from a pool of candidates and added to the hidden layer. It is called a cascade because the output from all neurons already in the network feed into new neurons. As new neurons are added to the hidden layer, the learning algorithm attempts to maximize the magnitude of the correlation between the new neuron’s output and the residual error of the network which we are trying to minimize.