DEEP LEARNING

With the aid and advancement of technology, we have witnessed it's application spreading across all domains and spheres of life. Artificial Intelligence, Machine Learning , IOT , Blockchain are considered as the next big thing in the world of technology and the same applies for Deep Learning. From snapchat filtering to image-voice recognition, and from medical image processing to stock market prediction, all these are the use cases of Deep Learning. 

Artificial intelligence, Machine Learning and Deep Learning are some of the buzzwords which hovers in everyone’s mind these days. Well,  you might know all these are interconnected but lets jump on to exact essence of the each:

ARTIFICIAL INTELLIGENCE:   AI implies getting  a computer to mimic human behavior in some way.

MACHINE LEARNING:ML is a subset of AI and it consists of the techniques that enables computers to figure things out from the data and deliver AI applications.

DEEP LEARNING: Deep Learning is a subset of ML that enables computers to solve the complex problems.

Lets deep dive into the crux of Deep Learning:

Deep Learning essentially and effectively revolves around Machine Learning technique in which a system takes the input in the form of images ,audio ,text through layers in order to predict the output.The main crux behind deep Learning is the way the human brain is able to filter information. Communicate with each other to give the final output. In the human brain, there are about millions and billions of neurons and each neuron is connected to  its neighbors. Essentially, that is what we’re trying to create, but in a way and at a level that works for machines.

AN INTUITION BEHIND DEEP LEARNING IS:

You get input from observation, you put your input into one layer that creates an output which in turn becomes the input for the next layer, and so on. This happens over and over until your final output signal!

Deep Learning Models can be classified as 

SUPERVISED

SEMI SUPERVISED

UNSUPERVISED

Supervised deep Learning is a function that maps an input and gives an output. It works  with labelled training data. Essentially, each example is a pair that is made up of an Input data(usually a vector) and the output value that you expect (supervisory signal). In a nutshell, it looks at stuff with labels and uses what it learns from labeled stuff to predict the labels of the other stuff.

For instance:

Same cycle is applied to classification of email as spam, recognising voices in an audio and  creating a model from speech to text.

SEMI SUPERVISED LEARNING:

Let's take an instance to throw some limelight on it.

- Your elders strive to imbibe a sense of inculcating good manners and moral values in you. At the same time, there are a volley of stuff which you learn by yourself by observing. This is a classic example of Semi Supervised Learning wherein semi supervised learning incorporates both -the labeled and unlabeled data for training. Semi supervised learning has been tried,tested and proven that is better than unsupervised learning. 

Unsupervised learning revolves around the relationships between elements in a dataset and able to give output without the usage of

There are scores of deep learning techniques, say Convolutional Neural Network, Recurrent Neural    Network,GAN etc

Lets understand main crux of DEEP LEARNING by application of CONVOLUTIONAL NEURAL NETWORK:

How our brain classifies an image!

We categorize things by recognizing the features. 

 In the above photo ,it's pretty difficult to decide if the person is a girl or an old woman

All the above images are addressed to understand that our brain functions on the features of the image it sees and then classifies it accordingly.In simple words, if a person is classifying an image, he/she will be able to classify it with the features of the image. It will read the each and every pixel of the image and then understand it

 

Deep learning Process

Deep neural network crystal clearly provides accuracy at an amazing level. They can learn automatically , without predefined knowledge explicitly through coding.

A deep neural network provides accuracy in many tasks.They can learn automatically, without predefined knowledge explicitly coded by the programmers.

 

MOST CLASSIC EXAMPLE

To grasp the idea of deep learning, imagine a family, with an infant and parents. The toddler points objects with his little finger and always says the word 'cat.' As its parents are concerned about his education, they keep telling him 'Yes, that is a cat' or 'No, that is not a cat.' The infant persists in pointing objects but becomes more accurate with 'cats.' The little kid, deep down, does not know why he can say it is a cat or not. He has just learned how to hierarchies complex features coming up with a cat by looking at the pet overall and continue to focus on details such as the tails or the nose before to make up his mind.

A neural network works quite the same. Each layer represents a deeper level of knowledge, i.e., the hierarchy of knowledge. A neural network with four layers will learn more complex features than with that with two layers.

The learning occurs in two phases.

The first phase consists of applying a nonlinear transformation of the input and creating a statistical model as output.

The second phase aims at improving the model with a mathematical method known as derivative.

 

Let's take an example to understand it comprehensively. Suppose a model wants to learn how to dance. After 20 minutes, it will be a random scribble.

 

After 48 hours of learning, the computer masters the art of dancing.

 

WHAT DOES THE ABOVE HUMAN MODEL. REFLECT??

 

Suppose you want to model to show a human is dancing. 

Now after  20min     ,you will find it's not completely trained. You will find only his hands moving without any other reflexive movement.

After 48 hours, you will find that the whole dance movement of the person who is trained  

This apparently reflects that Deep Learning takes time but the results are highly impressive in terms of accuracy of the model.

 

 

Convolutional neural networks (CNN)

CNN is a multi-layered neural network with a unique architecture designed to extract increasingly complex features of the data at each layer to determine the output. CNN's are well suited for perceptual tasks.

CNN is mostly used when there is an unstructured data set (e.g., images) and the practitioners need to extract information from it

For instance, if the task is to predict an image caption:

The CNN receives an image of let's say a cat, this image, in computer terms, is a collection of the pixels. Generally, one layer for the greyscale picture and three layers for a color picture.

During the feature learning (i.e., hidden layers), the network will identify unique features, for instance, the tail of the cat, the ear, etc.

When the network thoroughly learns how to recognize a picture, it can provide a probability for each image it knows. The label with the highest probability will become the prediction of the network.

Well, Deep Learning has spread its wings across all the spheres of the society. 

AI in BUSINESS:

Currently,AI is one of the most prudent means of customer demand and service management. From

 movie recommendation in Netflix to upgrade speech recognition in call-centre, Deep Learning has pl

From medical Image recognition to advanced movie recommendation. Fintech companies are growing even faster as they are using DL to save time and reduce cost.

FUTURE SCOPE:

The future of DL is endless and might go till eternity. 

- DL  with the aid of libraries such as Tensorflow might be integrated with Mobile Development .

- DL along with GPS will be really lethal 

- Self driving cars will become really common in future.

CONCLUSION:

Hence, Deep Learning Is Undeniably Mind-Blowing

Neural networks were invented in the 60s, but recent boosts in big data and computational power made them actually useful. A new discipline called "deep learning" arose and applied complex neural network architectures to model patterns in data more accurately than ever before.The future is DEEP LEARNING and the demand of it is at its peak in the market and will be rising for the next few decades.

It's the trend of the future and one should really learn, understand and apply it in real life.