Machine Learning & Deep Learning

Finding the Silver Lining? It’s right here!

The Crux of A.I.  ➳ Machine Learning

Let’s first understand the basics of Machine Learning & Deep Learning:

Machine Learning is a subset of Artificial Intelligence that includes abstruse statistical techniques, that enable machines to improve the tasks, based on experience.

Neural Networks

What do you understand by this technical term A.N.N.?

An Artificial Neural Network (A.N.N.) is a computational model, based on the structure and functions of Biological neural networks. Information that flows through the network, affects the structure of the ANN because a neural network changes – or learns, in a sense – based on that input and output.

Too technical? 😂

In simple terms, a neural network is a system of hardware and/or software patterned after the operation of neurons in the human brain. The first artificial neural network was invented in 1958 by psychologist Frank Rosenblatt, called Perceptron. It was intended to model how the human brain processed visual data and learned to recognise objects.

Artificial neural networks typically contain much fewer than the approximately 10¹¹ neurons that are in the human brain, and the artificial neurons, called units, are much simpler than their biological counterparts.

So in short, an ANN is an information processing paradigm, based on a collection of connected units called artificial neurons. As seen in the figure above, neurons are organised in layers: Input, Output and Hidden layers. However, the original goal of the neural network approach was to solve problems in the same way that a human brain would.

As of 2017, neural networks typically have a few thousands to a few million units and millions of connections. Moreover, their computing power is similar to a worm brain, several orders of magnitude simpler than a human brain. On the contrary, they can perform functions (e.g., playing chess) that are far beyond a worm’s capacity.

It consists of state-of-the-art learning algorithms, for example, Ranking web-pages, Spam email detection, etc. It is the field that is developing rapidly and improving new capabilities like WebSearch, PhotoTagging, etc. Some of the most common areas where ML is used on a large scale are listed below:

1. Data Mining
   • Web click data
   • Medical records
   • Biology
   • Engineering
2. Applications that cannot be programmed by hand
   • Autonomous helicopter
   • Handwriting recognition
   • NLP
   • Computer vision
3. Self-customising programs
   • Amazon/Netflix product recommendations
4. Understanding human learning
   • Real A.I.
   • Human brain

Arthur Samuel described it as: “the field of study that gives computers the ability to learn without being explicitly programmed.” This is an older, informal definition.

Tom Mitchell provides a more modern definition: “A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E.”

For Instance: playing checkers.

E = the experience of playing many games of checkers

T = the task of playing checkers.

P = the probability that the program will win the next game.

In general, any machine learning problem can be assigned to one of two broad classifications:

Supervised learning & Unsupervised learning.

Others: Reinforcement learning, Recommender systems