WHY DEEP NEURAL NETWORK: Difference between neural network and Deep neural network

 WHY DEEP NEURAL NETWORK: Difference between neural network and Deep neural network.

 Labeled data set is the prime requirement of Supervised learning. As a.  massive data is generated every day, AI Projects are innovated with rapid speed and choice of the algorithm is a critical issue:

Choice of the algorithm depends 

1.     Number of data sets (samples) and attributes  in that data set.(Dimensions of input and                output).

2.    Linearity and non-linearity in that data set

3   Performance.

Neural Network or Deep Learning: 

Algorithmic steps of neural network are 

1.   Need to decide Architecture of NN i.e.Number of hidden layers and number of units in that hidden layer.

2.    Apply training algorithm by Deciding hyperparameters like, Transfer function, No of iteration, learning rate etc to train the network.

3.    Test the performance

 

Major disadvantages are

Ø  If required performance is not achieved we need to repeat step 1 to 3 again.This cycle is time consuming as we need to change hidden layers ,units in the hiden layers and also hyperparameters and repeat the work. 

Ø  Some time  we need to change the learning law and repeat the work.

Ø  There is a possibility of over fitting or under fitting due to size of the data

and NN will not predict the correct result. 

It is time-consuming process and if NN will not working properly and need to choose another algorithm.Perticularly for very high level of performance two important things are 

1.  Need a lot of data. 

2. Able to train a big enough neural network i.e size of the neural network,a lot of hidden units, a lot of parameters, a lot of connections  in order to take advantage of the huge amount of data.

In above two cases performance of NN fail and Deep Learning will provide better results.Deep Learning will work better for large amount of data and large network 

Deep Neural Network Scale, both large amount of the data as well as size of the neural network ,a lot of hidden units ,a lot of parameters ,a lot of connections 

 

A deep neural network is simply a neural network with many layers.

The figure 1  shows a simple neural network and Deep Neural Network.The difference to a deep neural network (on the right) is clearly visible. The extra hidden layers provide a huge increase in computational power, which have allowed deep neural networks to reach amazing performance in multiple tasks.

 

 Figure1:Difference between Neural Network and Deep Neural Network

Advantages of using deep learning approach is its ability to extract features from data set by itself. In this approach, an algorithm scans the data to identify features which correlate and then combine them to promote faster learning without being told to do so explicitly. This ability helps data scientists to save a significant amount of work.

 Examples of Deep learning :Deep learning model learns to perform tasks directly from text,sound,image. It is used  for  high-end innovations like driverless cars, voice control in devices like tablets, smartphones, hands-free speakers etc and many more. 

ANN	Deep Neural Network
Suitable for small and medium size  data	Suitable for large size data
Less number of Hidden layers	Large number of hidden layers.i.e Deep layers
Do not identify features, Preprocessing is major task of feature extraction	Identify features in the network itself
Architecture consists of hidden layers and units	Architecture consists of convolution layer, Max Polling Layer and fully connected Layer
To identify face recognition image processing  techniques are used for  extract feature  and fed to Neural Network	To identify face recognition face image is directy fed to Deep Learning model