Series: Explaining machine learning to my mother (part 1) — Introduction
Why?
“How was your talk?”, my mother asked me over the phone.
“Was fine, I guess”, I replied wearily.
“What was it about?” she asks.
“It’s nothing you’d understand Amma”, I say.
“I’d still like to hear the name of the topic”, she pressed on.
“Bayesian Inference”, I concede, curtly.
“Oh! I have no idea what that means”, she says with a short laugh.
It was a tiring evening after a long day at work, I also had earlier that day, given a two and half hour talk on Bayesian Inference to a group of eager 20-year-olds. I simply wanted to stop talking.
The next day however the conversation came back to me, I was irritated with myself for running out of patience so quickly.
I also had a realization, my mother is one of the smartest people I know personally.
She could explain to a 13-year-old me — the history and current state of Indian politics, the intuition behind the theory of relativity and expound the essence of Bhagavad Gita — all in a single evening when she would find some respite.
If I am going to dismiss this person, with a single sentence “It’s nothing you’d understand, Amma” when asked to explain something as simple as Bayesian Inference, isn’t it time I truly examined, if I understand what I’m talking about?
“If you can’t explain it simply, you don’t understand it well enough.” — Albert Einstein
So, as a challenge to myself, to truly understand the work I do, and so that my mother who is forever curious about everything can learn about it, I am beginning this series on an intuitive explanation of machine learning — for Amma and her never-ending love for learning.
What to expect
This is a series of articles aimed at breaking down various pieces of machine learning in layman terms. The topics are arranged based on my idea of relevancy.
When necessary for the topic under discussion, I’ll try to explain the idea behind various machine learning techniques with as little complexity as possible. This doesn’t mean I’ll skip mathematics, because mathematics is probably the most fun part of machine learning.
I’m also challenging myself to explain the mathematics behind machine learning using simple intuition.
At a high level, I’ll be writing about the following topics in the following days (will continue to change/expand, I welcome suggestions).
- What is machine learning? Why is everyone talking about it?
- So, you predict something — How do you do that?
- Is it true that you can predict election results using posts from Twitter?
