An Introduction to Machine Learning

misc.Ria Ghosh

Machine Learning Robot

To begin with the buzzword “Machine Learning(ML)”, you will first need to focus on another word “Artificial Intelligence(AI)”. Yes, both are related. Machine learning is a part of AI.

The bigger term, AI is about creating an environment where you do not have to command the computer to do a certain thing, the computer can think and understand on its own.

Think of the air conditioner which turns off when there is no one in the room. It is programmed or designed in such a way that no external medium is required to turn it off. A lot of basic and everyday used technology are examples of AI. Like your phone adjusting the brightness of the screen according to the natural light.

Your cell phones and your personal computers are evolving every day. With the help of AI, they are becoming more intelligent and efficient.

In order to be flawless, one needs to remember and correct their flaws. Like, remembering your mistakes and not repeating them further or correcting them. Similarly, your cell phone which is a machine can now remember the words you use very often and predict them while you are typing a message or suggest you a new horror movie because you watch horror movies a lot. This amazing way of the machines to learn, adapt, improvise and improve themselves without explicitly feeding them programs to do the same is known as “machine learning”.

This amazing way of the machines to learn, adapt, improvise and improve themselves without explicitly feeding them programs to do the same is known as “machine learning”.

Now, the question is how the machines get to learn themselves. Just like your brain, which stores the information that you read in your books and use them when you are writing an exam. These machines store the data or information such as the words that you type and the number of times you use the word in a single sentence. This information is later used to predict the word that you want to use.

Similarly, the air conditioner records the temperature of the room at different hours throughout the day and adjust its temperature setting accordingly without you explicitly adjusting it with a remote.

It is true that now you don’t have to explicitly write a program to identify a person’s face, whenever he tries to unlock his phone. But there has to be some programming done to capture the facial features of a person in the beginning, After that those captured features must be stored as information which is later used to unlock the phone whenever needed.

Now, some method is used to match or compare the facial features stored, with the one that is currently available while unlocking the phone. These methods are known as algorithms which help in machine learning. These are categorized as

  • Supervised machine learning algorithms
  • Unsupervised machine learning algorithms
  • Semi-supervised machine learning algorithm
  • Reinforcement machine learning algorithms

We will discuss these algorithms in later articles.

Share and support us

Share on social media and help us reach more people