Source:pexels.com

As a father of 2 year old son, sometimes I wonder how am I going to explain my area of work to him as he grows up. I think many of you already had encountered such situations at home. Most of the kids use to think that Machine Learning is all about making cool robots, kind of Doremon who can do all of their work, specially homework 😉. So here is my attempt to simplify things a bit for better benefit of self as a renowned physicist Richard Fenynman once said:

If you can’t explain something to a child, there’s a chance you don’t understand it well and by coming to be able to explain it briefly and accurately, you’ll be learning.

Machine: First let us clarify what kind of machine we referring when we are saying, Machine Learning or in other words what kind of machine has the ability to learn. Generally computer scientists refer their computer systems as machines but it is not the case with the kids. Anything from a ceiling fan to TV , toy car to washing machine are machines for them. So we can define a learnable machine as anything which has a processor and a memory and a way to interact with it. Simple examples from their surroundings could be tablets, mobile , laptop or computer systems.

Learning : Thanks to our education system that kids and often parents also interpret memorisation for learning.So, lets make a distinction here, when Teacher asks you to learn Table of 2, he/she is basically asking you to memorise product 2 with first 10 natural numbers. Now lets understand learning, suppose teacher gives you a problem of guessing the basic algebraic operation in the below equation.

2 ……. 1 = 2

You can guess it as multiply(*) or divide(/). Now you are given one more equation.

2 ……. 1 = 2
2 ……. 2 = 4

You can still guess it as multiply(*) or exponent(^). In other words, even with two equations we are not able to learn the correct operation. Lets take one more equation then

2 ……. 1 = 2
2 ……. 2 = 4
2 ……. 3 = 6

Now , with the equations above, we are able to learn the correct operation multiply(*) . So, from the above examples we learnt to identify the pattern and now if we are asked to solve the following, we can easily do,

2 ……. 1 = 2
2 ……. 2 = 4
2 ……. 3 = 6
2 ……. 4 = ?

Not only machine learns predicting here but also with this much training machine has learnt Table of all numbers. Magic isn’t it 😎.

This is all about the fascinating terminology Machine Learning or ML. Identify the pattern from the given data and predicting for new data points. Also we saw the more the data, the better the learning.I hope it will satiate atleast a bit of curiosity about ML in the upcoming generation. Do share your views in comments.

M.Tech in Data Science, IITH, AI/ML enthusiast, part time shayar