Neural network

Imagine, you have to go to Russia But you don't know Russian language. you'll have no trouble finding your way, all thanks to google's real-time translation of Russian boards to English. this is just one of the several applications of neural network.let's talk about neural network.
Neural networks from the based of deep learning. a sub field of machine learning where the algorithms are inspired by the structure of the human brain. neural networks take in data train themselves to recognize patterns in this data and then predict the outputs for a new set of similar data. let's understand how this done.
                                                Let's make a neural network that differentiates between a square circle and triangle. neural networks are made up layers of neurons. these neurons are the core processing units of the network. first we have input layer which receives the input. the output layer predicts the final output. in between exist the hidden layer which perform most of the computations required by our network. 

Here's an image of a circle. this image is composed to 28 by 28 pixels. which make up 784 pixels. each pixel is fed as input to each neuron of first layer. neurons of one layer connected to neurons of the next layer through channels. each of these channels is assigned numerical value known as weight. the inputs are multiplied to the corresponding weights and their sum is sent as input to neurons in the hidden layer. each of these neurons are associated with a numerical value called the bias. which is then added to the input sum this value is then passed through a threshold function called activation function. the result of the activation function determines if the particular neurons will get activated or not. an activated neuron transmits data to neurons of the next layer over the channels. in this manner the data is propagated through the network. this is called forward propagation. in the output layer the neurons with the highest value fires and determines the output. the values are basically a probability. 

 For example here are near unassociated with the square has the highest probability hence that's the output predicted by the neural network. of course just by look at it, we know our neural network has made wrong prediction but how does the network figure this out. that our network is yet to be trained during this training process along with input. our network  also as the output fed to it.the predicted output is compared against the actual output to realize the error in prediction. the magnitude of error indicates how wrong we are in the sign suggests if our predicted values are higher or lover than expected. the arrows here give and indication of the direction and magnitude of change to reduce the error. this information is then transferred backward through our network. this is known as backward propagation. then based on this information the weights are adjusted. this cycle of forward propagation and backward propagation is actively performed with multiple inputs. this process continues until our weights are assigned such the network can predict the shapes correctly in most of the cases.

You might wonder how long this training process take. honestly neural network may take hours or even months to train. but time is a reasonable trade off when compared to its scope. let us look at some of the applications of neural networks. Facial Recognition cameras on smartphones in these days can estimate the age of the person based on their facial features. this is neural networks at play. first differentiating the face from background. then correlating the lines and spots on your face to a possible age. Forecasting neural networks are trained to understand the patterns and identify the possibility of rainfall or arise and stock prices with high accuracy. 

If you want to read articles about technology visit now