Why Machine Learning?
Machine Learning, AI, Deep Learning… these words need no introduction at the moment! We all have an opinion on, or sometimes even fantasize about AI taking over the world, thanks to numerous movies and TV series.
But what makes AI and Machine Learning such a buzz topic? Why are so many people interested in that? Here’s five reasons why I love ML so much and want to start a career on it!
My love for mathematics!
Ever since I was a kid, I loved mathematics. The reason was partly because I was good at it, and partly because it was so interesting. I would like to quote Neil deGrasse Tyson from his famous book, “Astrophysics for People in a Hurry”. He starts the book by saying “The universe is under no obligation to make sense to you!”. That’s true, but we have used our 3 pounds of grey matter in our head to invent Mathematics. And Math is the language of the universe. All of our accomplishments as a collective human race wouldn’t be possible without our understanding of mathematics.
With no surprises, every concept in Machine Learning has it’s deep roots in the concepts of mathematics. In fact, every other algorithm in computer science is actually a subset of the ocean, that is, mathematics.
Advent of Data
In the 50,000 to 300,000 year old human civilization, data or information has never been readily available ever, as we have it now. We have managed to generate more data in the last half a century, than we have in the previous 50 centuries. There is not only such huge amounts of data, there is also huge variety of data.
The real question is, what have we managed to do with so much information? That’s when Data science and Machine Learning comes into picture. With so much data and so many different types of data, it was only a matter of time before we began to think about doing something useful with it. And the insights we get from the data, has helped us make informed decisions in so many crucial places.
Cricket(Sports) and Data Science
I live in India, and here there are very few things which has the same influence on society as Cricket. The amount of people who sit and watch a World Cup match in India is easily in the billions. Cricket sure did have an influence on me. I grew up watching cricket and even imagined being part of the national team, although that remained a dream for me.
I took up computer science as my under grad, and around the time I was in my first year in college, I got to know about a really cool job. I came to know that there is a data scientist as part of every IPL team, who analyses players’ data and comes up with a game plan or strategy to improve the team. This was a revelation for me, as I thought it was only the coaches, support staff and the captain who make decisions. Little research got me to realise that there is a data scientist for the Indian Cricket team as well, and data scientists actually play a crucial role in every sports team in the world. This amplified my interest in Machine Learning as I could combine two things I love and start a career with that. How cool!
Possibilities are endless
Once I was into Data Science, I started exploring the various algorithms and techniques. One thing that struck me the most was, there was no fixed algorithm for a particular domain. The number of parameters you can tune to improve your algorithm is endless. This gives us a playground to explore so many endless possibilities given the amount of data we have. You want to improve your accuracy? Try getting more data. You cannot get more data? Try amplifying the data you already have. You think this algorithm is not fitting the data properly? Try a different one. You think this is the best algorithm that can fit this data? Try tuning the parameters to improve it. This gives an excitement to building algorithms.
There are also endless possibilities in the domains you can apply machine learning. Since data is omnipresent, and in such huge volumes, machine learning cannot be restricted to any particular part of our lives. This has led to humans using machine learning in sports, weather forecast, stock market prediction, AI based robots and machines, health care, Image recognition, Speech recognition etc. The world is your oyster!
…Cause it’s really interesting!
The reason why machine learning is a buzz word these days is just because it is so damn interesting. And for me, the most interesting part of machine learning has been Image Processing! Here, we segue into Deep Learning. The various algorithms in Deep Learning tries to mimic the way the human brain works.
We all learn from our experiences. Deep learning algorithms are no different. Given enough examples of a particular class of images, the algorithm can learn various features of the image and predict a completely new image under the same class. If we understand linear algebra, the underlying mathematics behind this is pretty easy, but the fact that we somehow applied that to images, to actually build various algorithms is very interesting and fascinating. And trust me, image processing has actually taken over our everyday lives. Facial recognition systems are available in our phones, our offices, and it is used in crime investigations (we have all seen in movies :P). Moreover, the insights we get from analysing data and the process we follow to get there, is something very engaging.
And there we go. That’s some of the reasons I love Data Science and Machine Learning!
I believe ML and AI is gonna rule the future, as we have still not mined all the data and information we produce. And as a consequence, there will be demand for more data scientists in the near future. I hope to see huge advancements in Machine Learning in the next decade!
Here are some useful courses to begin your Machine Learning journey :
- Machine Learning, Andrew Ng — Coursera
- Deep Learning Specialization, Andrew Ng — Coursera
- Machine Learning A-Z : Hands-On Python & R In Data Science — Udemy
Happy learning! :D
