Deep Learning Simplified: For beginners

# Deep Learning Simplified: For beginners

Deep Learning Simplified: For beginners

In today’s world full of intelligent software, an app can click pictures and tag names automatically. A popular example is the Google Photos app that most of us use on a daily basis. Do we even understand what it does and what is the exact algorithm behind it that is executing continuously? Well, it is not compulsory to understand complex mathematics or end-user. They just use the application and enjoy automation. It is the responsibility of software developers to dig in the science, mathematics, and algorithms running behind the automated functionality.  If you aspire to learn the techniques behind such applications Machine learning technology is a solution. In this wide technical domain, deep learning is an expertized methodology that allows engineers to design, develop, and test such applications in the real world.

Here, interesting thing is,

• Headphones/earphones are not only used for listening, but they are also used for real-time translation.
• Scanners are able to analyze people’s images and differentiate one face from another in the group picture.

Do you want to learn the engineering behind such apps?

Welcome to the world of Deep Learning.

Deep Learning is at the core of many front line Artificial Intelligence (AI) applications.

The purpose of this article is to explain deep learning in simple words so that confusion of beginners or anybody who wants basic will be out of the scary sea of confusion. This article does not say that maths and code is bad but the idea here is to understand the concept first and clear your thoughts. Once one is clear on this conceptual side, then gradually one can progress towards, required maths, code, and software libraries that are available.

Neural Networks

Neural Networks is a set of techniques intended to study the system that is used by the human brain for its working.

The biological neurons motivate the architecture and the function of the neural networks.

Every artificial neuron is demonstrated as:

• Every neuron accepts inputs
• Make Addition of weights and biases to the received input
• Make Sum of inputs with weights and bias
• This process activates the Activation function
• An activation function is responsible for taking the weighted sum of the preceding stage and produce the output as a later stage