A Baby is born - Supervised Learning
Machine Learning can be broadly divided into two distinct set of Learning Algorithms — Supervised Learning & Unsupervised Learning.
In this article, we would be understanding Supervised Learning with real life analogies. Let’s get started!
A Baby is born!
The Baby is the threshold of innocence & ignorance for a human being.
A Baby is born with nothing but unconditioned reflexes.
The baby knows how to breathe, how to sleep.
The baby knows eyes should be constricted when illuminated by light!
The baby starts salivating on tasting milk!
If disturbed from sleep, the Baby’s instinct tells the baby to cry!
As Baby grows up, Parents try and teach Baby how to crawl, walk, talk and recognize ! A new born baby doesn’t know what is meant by mommy, daddy, nor does it know whom it’s addressed to!
“This is daddy. Say daddy!
D-A-D-D-Y!
This is Mommy. Say Mommy!
M-O-M-M-Y!”
.
.
.
.
After weeks or months of training through lullabies, the baby can finally point out when asked :
Where’s Daddy?
Where’s Mommy?
The Baby also develops other abilities in the mean time. Baby starts recognising voices of it’s mom & dad!
Then after 2 long years, with newly grown teeth, the baby is finally able to say those first words which every parents long to hear!
The Baby, after being taught for two years through lullabies , can finally say “Daddy” “Mommy”, can recognise them and even their voice!
The Baby is now Naturally intelligent!
Now, let’s take the example of your computer.
It has some unconditioned reflexes, just like the baby!
It reacts to clicks.
It performs arithmetic and logical operations
It shuts down on low power!
Now, you want to teach your computer ( Baby ) to recognise people (mommy & daddy).
This time, instead of verbally teaching the baby with lullabies & dance, you decide to use a table to teach computer to do the same :
We keep adding rows to the table, with each row having a different image of Mommy and Daddy, at different angles. We feed this table to our computer and make it learn the table with help of an Algorithm, so that next time the computer sees my face, it knows “daddy is here”!
The Computer is now Artificially Intelligent !
.
.
Here, Algorithm is just a procedure to make someone learn something, just like a Baby’s parents might have resorted to dancing or singing lullabies or something that might catch the Baby’s attention, in order to make the baby learn “Mommy” “Daddy”.
Just like continuously sitting in front of a baby with a straight serious face, trying to teach “Mommy” “Daddy” won’t work, similarly, just feeding the computer table of pictures won’t!
Parents all across the world resort to their own unique techniques to teach the Baby. Depending on parental skills and the Baby’s inherited intelligence, some learn quickly, some take time.
Similarly, us computer users, will resort to some unique way to teach our computers. Depending on our teaching abilities & our computer’s configurations, some computers will learn better and faster compared to others!
But, here is the catch!
A baby would take at least a year of teaching to recognise “mom” & “dad” . With the help of Supervised Algorithms, your computer will take few minutes to do the same!!
That right there is the basic understanding of Machine Learning in general, Supervised Learning in particular!
To sum it up, Supervised Learning is the process by which computers are able to make decisions with the help of “labelled data” !
Now, if you are still reading, here’s something extra :
The word labelled data is very important. We are labelling my picture to “Daddy” and my wife’s picture to “Mommy”, just like the parents labelled themselves to the sound “M-O-M-M-Y” and “D-A-D-D-Y”
We can perform Supervised Learning only and only on labelled data!
We would be seeing the relevance of Labelled data and it’s other variant (Unlabelled data ) as we move forward in our journey to becoming an AI literate!
