Dear Ranpreet Kaur,

AI means getting a computer to mimic human behavior in some way.

Machine learning is a subset of AI, and it consists of the techniques that enable computers to figure things out from the data and deliver AI applications.

Deep learning, meanwhile, is a subset of machine learning that enables computers to solve more complex problems.

Artificial intelligence is a broader concept than machine learning, which addresses the use of computers to mimic the cognitive functions of humans. When machines carry out tasks based on algorithms in an “intelligent” manner, that is AI. Machine learning is a subset of AI and focuses on the ability of machines to receive a set of data and learn for themselves, changing algorithms as they learn more about the information they are processing. 

Deep learning goes yet another level deeper and can be considered a subset of machine learning. The concept of deep learning is sometimes just referred to as "deep neural networks," referring to the many layers involved. A neural network may only have a single layer of data, while a deep neural network has two or more. The layers can be seen as a nested hierarchy of related concepts or decision trees. The answer to one question leads to a set of deeper related questions.

Thank you for answer Mr.Milon. One more doubt I want to ask.

Like Artificial neural network, are there any algorithms of deep learning or it itself is single approach?

You may find useful information for a better understanding the way these artificial networks work in the online book at the address:

http://neuralnetworksanddeeplearning.com/index.html

Deep learning is a very broad term - it's a special sub-field of machine learning. There is no single approach, it encompasses various techniques, models and algorithms.

Artificial neural networks (ANNs) are just a tool, one of many used in machine learning. Again, it's a very broad term as there are plenty of different types of ANNs and algorithms to train them.

Explaining this relationship to my students, I would make a following analogy to the world of mathematics: if machine learning is calculus (a very broad term), then deep learning is analysis (some more specific but still pretty broad branch) and ANNs are functions (used throughout both branches along with many other concepts, with probably more importance in the latter).

ANN is the broader term between the two. In fact, any DL model is an ANN but the opposite is not quite true. DL refers to recent developments (~20y) that lead to NN with multiple layers (in a hierarchical fashion) while ANN refers to any biologically inspired NN.

The difference between neural networks and deep learning lies in the depth of the model. Deep learning is a phrase used for complex neural networks. The complexity is attributed by elaborate patterns of how information can flow throughout the model.

Deep Learning  = Artificial Neural Networks + more of everything

Both terms relate to Machine Learning which is a subfield of Computer Scienceand Artificial Intelligence. The term Artificial Neural Networks (ANN) refers to a family of machine learning algorithms. The definition is not strict at all and fits usually for any trainable algorithm using a network like a topology.That said, most people would probably imply a multi-layer perceptron (MLP) neural network with a sigmoid activation function as they have been used the most.

It can vary from case to case, but usually, ann's are trained on handmade features and not on raw input. I'm referring to the feature extraction being a separate part of such a system in some cases. Furthermore, especially early neural networks used only a single or a few hidden layers and were not able to train on very abstract features. These universal function approximators had their limits. Since 2012 however, we have made significant progress. Suddenly we are now able to train such neural networks on a dozen layers and more. With massive computing power and huge amounts of data these algorithms were now able to do the feature extraction themselves hence work on raw and very abstract data. This new way of using the same algorithms has led to the term of Deep Learning and it usually implies the use of huge amounts of raw input data, high level of abstractions and massive computations. It is, however, not limited to a specific neural network technique but to many very different ones. It has become more of a buzzword as it leaked more and more into general media as it encompasses nowadays every type of modern learning algorithm.

In the simplest form, deep learning deals with a myriad of layers such as fuzzy and convolutional layers which adds new capabilities to an ANN like image, sound recognition. These layers aim to reduce the dimension of information into the vector scales to glean and extract patterns or salient features.

I explained this in other forums (ref: https://www.researchgate.net/post/AI_vs_Machine_Learning_vs_Artificial_Neural_Network_vs_Deep_Learning), and will summarize below

Artificial Intelligence (AI): branch of computer science that seeks to create intelligent systems through diverse avenues such as knowledge gathering, reasoning, perception and learning of prior assertions

Machine Learning (ML): process of applying Artificial Intelligence (AI) in a way to provide systems with the ability to learn and improve on prior experience towards performing a given task. Examples of ML includes Decision Trees (DT), Support Vector Machine (SVM), Artificial Neural Network (ANN), K-Nearest Neighbours’ (K-NN) and Regression Analysis (RA)

Artificial Neural Network (ANN): refers to a type of Machine Learning (ML) that tends to simulate biological neural networks, and uses connected nodes (that models biological brain neurons). Examples of ANN includes Bayesian network, Auto Associative Network (AAN), Back Propagation Network (BPN), Feed Forward Network (FFN) [such as extreme Learning Machine (XLM), Single Layer Perceptron (SLP) and Multi-Layer Perceptron (MLP)]

Deep Neural Network (DNN): a variation of Artificial Neural Network (ANN), however with consideration for more dense hidden layers. The process of training DNN for learning is typically termed Deep Learning (DL). Frequently used DNNs include Convolution Neural Networks (CNN), Recurrent Neural Network (RNN) and Deep Belief Nets (DBN)

Convolution Neural Networks (CNN): a type of Deep Neural Network (DNN) where each neuron in the network layers has a two or high dimensional filter which is convolved with the input of current or previous layer.