What’s Machine Learning?

In the previous post by Vedant Bahel on Artificial Intelligence, you were introduced to Machine Learning. Now, let’s have a look at what exactly it is.

Machine Learning has become a buzzword these days, something which is an integral part of of our lives and we use it in almost every domain we work in. Even in fields that we feel are completely irrelevant to this technology like music and art, whereas industries like Fast Moving Consumer Goods  (FMCG), Transportation, Healthcare, Automobile, Education, and especially banking and computer industries which have inculcated this technology into almost all of their operations and technologies, right from top to bottom.

Being such an important technology it becomes necessary for all of us to have the knowledge of the field because our future relies on this.

So, let’s get ready and dive deep into the ocean of machine learning.

But what is machine learning?

According to the definition, machine learning can be defined as

“The field of study that gives the computer programs the ability to learn without being explicitly programmed is called Machine Learning. – Arthur Samuel

” A computer program is said to learn from an experience E with respect to some class of tasks T and performance measure P, if its performance at tasks T, with respect to P improves with E.” – Tom Mitchell

To break it down into simple words.

Suppose we have a program whose task is to play a game of chess, so here

T: Task of Playing Chess

E: The experience of playing many games of Chess

P: Probability that the program will win the next game of Chess

Thus, according to the second definition, if the probability of the program to win the next game of chess improves with its experience of the task of playing chess. To put it simply, if the program gets better at playing chess with every game it plays we can say that the machine is learning to play chess, or perform the task with greater accuracy than before as it gains experience in performing the task.

We are intellects who created machines, how can they beat us?

Well, this is just like what happens with humans, we get better at a particular job if we perform a task repeatedly.

But at the end of the day, we have physical limitations and not enough patience to perform the same task for 100,000 times, which machines can do. Thus they can gather a great amount of experience of a task in an extremely short period of time and get super accurate at performing the task assigned that is probably out of our scope.

Which is why these programs can study a humongous amount of data, identify the trends in the data and use them to predict the final results of any other relative random values that we give as an input to our program, everything in a short period of time. The performance keeps improving with improvement in technology and increased computing power.

To conclude, Machine Learning is all about training a program, that uses a particular algorithm (that we’ll learn about in future posts) on lots of data beforehand so that it gains some experience (much like practicing for a sport) and does a great job predicting the output values of random input values in the future(the final game).