Deep learning basic concept for beginners.

                                      In this blog, we will study some basics of deep learning.

Machine learning is an Application of Artificial intelligence and deep learning is a class of machine learning.

Table of contents

1.  history of deep learning.

1. McCulloch and pitts.

2.  Frank rosenblatt.

2.  Function of perceptron.

3.  Basics 

1. What is meant by neurons.

2.  What is meant by neural networks (biological).

3. What is the meaning of artificial neural networks?

4.  What is meant by deep learning and Types of deep learning.

1. Conventional neural network.

2. Recurrent neural network.

1. History of deep learning

a. McCulloch and pitts

               It all started back in 1943, when An American neuro physiologist Warren McCulloch and logician walter pitts created a computer model based on human brain neural networks.

     A classical paper by Warren McCulloch and Walter pitts "A logical calculus of ideas immanent in nervous activity" played a very important role in the process of deep learning because it helped in the development of brain theory.

b. Frank Rosenblatt.

                           Frank Rosenblatt was an American psychologist and researcher at Cornell aeronautical laboratory. He is known for Perceptron. Perceptron was invented by Frank Rosenblatt.

2. What is meant by Perceptron? and function of Perceptron.

                         Perceptron is an algorithm designed to recognise the pattern. It gives ability to the machine for recognising the patterns.

3. Basics 

                   Before studying deep learning, we need to understand some basics, these basic terms will help us to know more about deep learning.

1. neurons.

2. neural network (biological)

3. artificial neural network.

A. Neuron.

Brain neuron

                             Neuron is an individual or basic working unit of the brain, it is a specialised cell whose primary function is to transmit the information through electrical and chemical signals. Neurons are also called nerve cells.

 Neurons are found in the brain, spinal cord and peripheral nervous system (PNS). On an average humans have around 86 billion neurons and An elephant have around 257 billion neurons.

B. Neural network.

Biological neural network

                            Neural network is interconnection of neurons means one neuron is connected to another. 

A biological neural network is made up of biological Neurons.

C. Artificial neural network (ANN).

                        Artificial neural networks are decorated the same as human or biological Neurons. Artificial neural networks require both hardware and software for working.

 Artificial neural networks are made up of artificial neurons, which function to mimic the human brain.

4. What is meant by deep learning.

                       Deep learning deals with artificial neural network algorithms which are inspired by the structure and function of the human brain. Deep learning is based upon artificial neural networks (ANN).

Types of deep learning.

1.  conventional neural network.

2.  recurrent neural network.

A. Conventional neural network.

                                                   Conventional neural network is designed for recognising the image or pattern. CNN algorithms are used for pattern recognition. CNN algorithms need to be trained before recognising the pattern.

 Conventional neural network is also known as shift invariant or space invariant.

Example : if the CNN algorithms were trained to recognise two breed dogs like German Shepherd and Labrador retriever and if you want CNN to recognise a bulldog, it will try to recognise it depending on the shape and edge of an object and bull dog is also a dog breed, so sometimes CNN will recognise it.

  But, if you want a conventional neural network to recognise a cat it will not recognise it because it is only trained to recognise dog images.

 CNN algorithms need some input data or training before recognising an image.

B. Recurrent neural network (RNN).

                                 Recurrent neural network is designed for predicting the sequence. RNN works with sequence prediction problems. It's function is to predict the upcoming value in a sequence.

Example : A, B, C, D, E the next in the sequence is F. Although it is just an example recurrent neural network is designed for predicting big data or big values in a sequence.

Summary.

                     In this blog, we have learnt the history of deep learning like when it started? By whom it started?. We have studied what is meant by neurons, neural networks and artificial neural networks. We also learnt the meaning and types of deep learning.