Artificial Intelligence – A brief Introduction and history of Deep Learning

In this article we will discuss about Artificial Intelligence and its applications. We all know that Computer is a device that can store, process data and exactly does what is instructed in the program.

If you save a picture in computer, then it knows only that it is JPEG or PNG image. It doesn’t know what is in the picture. You cannot tell your computer to get all of my photos unless you name the files with your name and you search for “name*.png”.

“A computer is a stupid machine with the ability to do incredibly smart things, while computer programmers are smart people with the ability to do incredibly stupid things.”

― Bill Bryson, I’m a Stranger Here Myself: Notes on Returning to America After Twenty Years Away

But not anymore, when the programmers are smart, computer can also become smarter.

The problem here is we want computers to think like us.

What is Artificial Intelligence (AI)?

Artificial Intelligence (AI) is a technique which simulates the human intelligence in machines using the programs that mimic human thinking and actions.

There are many applications of AI that made computers to think like a human.

Source: Devskrol.com – Image by Author

Machine Learning:

Machine Learning is a subset of Artificial Intelligence. It is a system that learns from the past data with the help of algorithms by finding patterns statistically and without programming explicitly.

As it is learning from the past experiences, most of the Machine Learning algorithms can replace the expert’s system.

For example, A real estate businessman can predict the price of a property based on certain features using the knowledge he gained from his experience. A machine learning algorithm can handle this same job if the real estate business man’s experiences are given to the algorithm as a dataset in a proper way.

In the same way there are a few other type of machine learning techniques that provides the ability to identify objects/text/digits from a picture or identify a person from the picture. But these techniques are not a replacement of an expert’s system. It is a common task or thinking of a human.

So, these tasks need more complex programs that should work like a human brain.

Neural Network:

In 1943, A neurophysiologist Warren McCulloch and a young mathematician Walter Pitts published a paper “A logical calculus of the ideas immanent in nervous activity”. They examined and understood how brain works and came up with a highly simplified model of an artificial neuron.

This model contains a layer of one neuron which has an input layer and an output layer. There is an threshold function which is applied after the processing of neuron, which controls the output of the neuron to have an output in binary format, i.e. to know whether that neuron is ON (threshold of one) or OFF (threshold of zero).

This Neural Net is called a perceptron which can only feed forward. i.e. the data passes only in one way. (If you are aware of backpropagation, this Neural Net has no ways of back propagation.)

This model can find a simple pattern from the data.

Deep Learning:

After the on and off research about Neural Net, we have got more complex Neural Network which has more than one hidden layer of neurons and each layer can have more than one neuron.

In each layer, the dimension of the network grows as the number of layers increases. As the network goes deeper, the learning goes deeper. Thus, this network is called Deep Neural Network and this learning is called Deep Learning.

Natural Language Processing (NLP):

NLP is another application of artificial intelligence, which processes the human languages such as speech or text. Common example of NLP is auto completion of sentences (example: Google search), Chat bots, speech to text conversion etc.

Computer Vision:

Computer Vision is another field of AI where the computers programs are trained to have a view like a human. Using the deep learning models, objects can be identified from the pictures or videos. A famous example is a self-driving car, which detects objects just like a human and drives the car accordingly.

Robotics:

Robotics is not only an AI but also a combination of machines, science, engineering and technology. Robotics mainly focus on hardware that does actions like a human. Examples are Robots that can reach and work on places where man cannot such as space missions, research in Antarctic, Underwater exploration etc. It also plays a major role in healthcare. Example: The da Vinci® Surgical Robot, The PARO Therapeutic Robot etc.

Source: https://www.meme-arsenal.com/en/create/meme/1470239

A brief history of Deep learning:

Here is an infographic of a brief history of Deep Learning. Even though we cannot show all the events in one single picture, I have added the most important events here, according to me.

Source: Devskrol.com – Image by Author

Conclusion:

In this article, we have discussed about the important applications of Artificial Intelligence (AI) and a brief introduction to those applications.

We have also seen an infographic about the history of Deep Learning.

We will look deeper in all the above topics in next articles.

Thank you for reading our article and hope you enjoyed it.

Like to support? Just click the like button ❤️.

Happy Learning!🎈

Thanks to:

http://www.mind.ilstu.edu/curriculum/modOverview.php?modGUI=212

https://news.mit.edu/2017/explained-neural-networks-deep-learning-0414

https://www.dataversity.net/brief-history-artificial-intelligence/

https://en.wikipedia.org/wiki/Artificial_neural_network

https://en.wikipedia.org/wiki/AI_winter

0

2 Replies to “Artificial Intelligence – A brief Introduction and history of Deep Learning”

Leave a Reply

Your email address will not be published. Required fields are marked *