Deep Learning: Not very Deep

Deep Learning a well known term to everyone in this decade — it is a surprisingly very powerful technology to use. Many people thought that Deep learning is very hard to understand. But luckily they are wrong.

Deep learning depends on neural net heavily and these neural nets are the replica of human brain. So if you understand the human brain,you can easily learn the deep Learning.

So DEEP LEARNING is a subset of machine learning and artificial intelligence which imitates the working of the human brain to process the data and identify the pattern in data.

For a 3d simulation you can checkout this youtube video:

Let’s check some applications of deep learning. DL is nowadays applied in various major public sectors like healthcare, Face recognition, Cyber security, Translators, NLP.

A great example how can deep learning techniques used to generate memes.

Check the article above.

Now Learn about the tools used in doing Deep Learning. The primary tools are used in deep learning is a programming language which can be chosen from a various range from R to python.But nowadays the Python is gained more popularity due to great frame works in this language.

The frameworks are tensorflow created and maintained by google,Theano,pytorch awesome deep learning libraries.

According to the tensorflow website:

“Whether you’re an expert or a beginner, TensorFlow is an end-to-end platform that makes it easy for you to build and deploy ML models.”

Image source:google & quora

Now the keras created by François Chollet which helps the users to build the neural networks very easily and deploy them in the production with ease. This is like the front end wrapper which makes to do deep learning coding with the tensorflow like a sugar quoted candy.

You can create a deep learning model with keras just about 5 lines of code compare to hardcode every single neural network connection in pure tensorflow.

Where in tensorflow it quite long:

Keras make it easy to get managed model very quickly and serve them to the cloud.

But both have advantages and disadvantages.Keras make things more fast where tensorflow give you more control over to build your network.

You can try these and discover the world of deep learning.

A sample code snippet using keras and tensorflow:

That’s all for today.