Latest news about Bitcoin and all cryptocurrencies. Your daily crypto news habit.
Every day, companies gather a bunch of data. Now they want to take advantage of those data. Thatâs why machine learning (ML), artificial intelligence (AI) and more widely data science domains are rising since quite a few years.
I work as a PhD student on a thesis to build a model of drones systems to check for failures.
Making models involve mathematics, also Iâve got plenty of data to analyze during experiments. As machine learning and modeling systems are very close, I decided to learn ML.
There are plenty of resources and specific vocabulary so it seems a bit hard to find your way. In this article, Iâll introduce you to machine learning and explain to you why it worths it to study it.
Is Machine Learning useful for real?
Even though you donât realize it, weâre surrounded by ML systems in our daily life. As soon as you go online to do researches you trigger ML algorithm made by advertising companies.
That means if youâre looking for a âcute small black and white dogâ on the internet, every website youâll visit after that will now try to sell you âsuper awesome cute small black and white dogsâ.
Besides, when you receive an email your inbox will detect if its a spam or a regular email.
Go boardgame, a Chinese strategy board game for two players
Did you heard about AlphaGo becoming the best Go player in the world? Its ability to beat Go master comes from training by watching how masters play during competitions.
Thanks to big data analysis, machine learning can help us to find the rhythm (or pattern) in the data.
Analysis results will improve your understanding of the data. Then you can use those insights to be better at doing something and make the most appropriate choices!
People trying to define what is Machine Learning
Letâs skip the part where I copy-paste the ML definition from Wikipedia. Instead, I prefer to copy-paste quotes from influent people in ML domain who try to define it.
Arthur Samuel: ML is a field of study that gives computers the ability to learn without being explicitly learned.Dr Tom Mitchell: A computer program is said to learn from experience E with respect to some task T and some performance measure P. If its performance on T, as measured by P, improves with experience E.
Theyâre both the first man at doing something. A.Samuel is the first man who developed the first successful self-learning program and Dr Michell is Chair of the Machine Learning Department within the School of Computer Science.
Well, those definitions are not that easy to grasp. To do so, Iâll give you an example of a dog, its name is Garfield.
Garfield trying to figure out what âgo get the ballâ means
Garfield is young but also the lazy kind of dog. You tried many times to throw away a tennis ball and ask it to get it back but it didnât.
To teach it to get the tennis ball, weâll give it some food whenever it comes back with the ball. After repeating this exercise a couple of time itâll begin to learn. It wonât understand what âgo get the ballâ means, but itâll know it can get food if it gets the ball and comes back with it.
Now go back to Dr Mitchell definition of machine learning and try to guess what are T, E and P in this example.
Actually, in this little history, the task is Garfield going to get the tennis ball when you ask it to. The experience is when you trained it with food, the performance is the probability for Garfield to come back with the ball.
Letâs wrap things up:
T: the task to performE: become better at T by doing itP: how are you good at doing T
A dog needs a lot of practice to develop its new abilities. The same applies to computers and ML.
The machine will give more accurate answers after practicing for a long time. Thus, itâs decisive to provide it with a bunch of good quality data.
Save your time, make computers work for you
Fine, ML do the same thing as a child in school: learn. Why do you need to use ML if its only purpose is to learn? Actually, machine learning programs donât only learn what we teach it faster than a child. It also analyses data to come up with insights and give the most accurate answers.
Garfield doesnât need to understand English, what you ask it is to behave properly in a situation: you throwing a ball and yelling at it.
Imagine youâre a successful sports car seller. To become richer, you want to sell your cars as expensive as possible but not cross the limit, so people still buy it.
You have data about past sells but figuring out insights from it by yourself might be tiresome. Instead, you can feed you machine learning program with your data without looking at your data. Your ML program will figure out what makes the prices of cars and come up with a formula.
Of course, you can do it all by yourself and find out the formula without using a computer. Yet, itâll drain your time without having a guarantee about finding the most accurate result.
Can Machine Learning do anything else ?
You can apply ML to almost any domain. After looking a bit about the variety of problems it can solve, I found a surprising one.
In this article from UW Data Scientist, the author suggests you to use ML to find a girlfriend. Here is one of it charts showing characteristics weights.
After answering some questions about you, its algorithm will give you the probability for you to find a girlfriend.
Thatâs it, your introduction to machine learning is over. I hope you enjoyed reading it and it helped you to find out what is ML.
See you soon for a next article about the different kinds of ML. Please let me know if youâre interested in it, donât forget to give this article some đ claps to support me.
Have a good day!
Brief Machine Learning Introduction using Dogs and a Tennis Ball was originally published in Hacker Noon on Medium, where people are continuing the conversation by highlighting and responding to this story.
Disclaimer
The views and opinions expressed in this article are solely those of the authors and do not reflect the views of Bitcoin Insider. Every investment and trading move involves risk - this is especially true for cryptocurrencies given their volatility. We strongly advise our readers to conduct their own research when making a decision.