Artificial Neural Network - what can artificial neural network do

                                            Artificial neural network

Artificial neural network is an interconnected group of nodes. it is one of the main tools used in machine learning algorithm. “neural” name suggested from the brain system . interconnected group of units or nodes represents an artificial linked neurons, blue lines represents in the given figure below the connection from the output of one artificial neuron to the input of another.

Let the input is called I1,I2,I3 and the hidden as H1,H2,H3,H4,H5,H6 and output as O and W(I1H1) is the weight of linkage between I1 and H1 nodes.

Following frameworks in which ANN(Artificial Neural Network) depends---

• Assign Random weights to all the linkages to start the algorithm

• Using the inputs and the (Input ->Hidden node) linkages find the activation rate of Hidden Nodes 

• Find the error rate at the output node and recalibrate all the linkages between Hidden Nodes and Output Nodes

 • Repeat the process till the convergence criterion is met

Warren Sturgis McCulloch and Walter Harry Pitts created a computational model for neural networks based on algorithms and mathematics in 1943.This  Algorithms called  threshold logic. This neural network divided into two approaches. One approach focused on biological process in the brain while another approach focused on artificial intelligence neural network. Neural networks require data to learn.

Components of an artificial neural network :-

 A neural with label j receiving an input P_j (t) neurons consists of the following components:

an activation a_j (t), depend on discrete parameter

a threshold θ_j , it keeps fixed unless changed by a learning function

an activation function F which computes the new activation at a given time t + 1 from a_j (t), θj and net input P_j (t) give rise to relation is:-

a_j (t+1) = f{a_j (t), P_j (t), θ_j} .

an output function F_out computing the output from the activation

o_j (t) = F_out {a_j (t)}.

Output function is known as identify function

Input neuron serves as input interface for the whole network and the output neuron serve as output interface of the whole network.

Connections:- the network consist of connection each and every neuron output i connected to input j . Each network assigned a weight w_ij.

Propagation function :- This function compute the input P_j (t) from the neuron o_i (t). so, j is the successor of i and i is the predecessor of j.

P_j (t) = ∑ o_j (t) w_ij

             i

Learning rule :- rule of learning algorithm which modify the parameter of neural network. The learning technique is to modifying the weights and thresholds of the variables within the network.

Different types of neural network :-

1. feedforward neural network

2. recurrent neural network

3. convolutional neural networks

4. Boltzmann machine networks

5. Hopfield networks

Task that can neural network perform :-

1.Gnerating CGI (Computer Generated Imagery) faces

2. Machine translation

3.making car drive automatically on the road

4.Reading our minds

5.Fraud detection

And some other tasks which can neural network perform perfectly