Artificial Intelligence, Machine learning and Data Science

Of late we have been surrounded by the terms 'artificial intelligence, 'machine learning' and 'data science'.

All the leading economies around the globe, have already begun investing in this domain. Extensive research work in this direction has started to become ubiquitous. But the question is, what are these terms exactly, do they bear any relation with each other and why is everyone discussing about them?

Machine learning and data analytics, essentially form the backbone of artificial intelligence. Simply put, artificial intelligence is the ability of your machine to think independently, without the intervention of the user and produce results. The machine needs to taught first, only then can it start thinking and produce results. But, the interesting point to ponder over is, how can we teach our machine to think, to produce results. Well, this can be achieved using machine learning and data analytics. The underlying principle is, teach your machine (machine learning) using the past data (data analytics) to make your machine think (artificial intelligence).

In machine learning, we make our machine "learn" based on various algorithms and the past data. The machine uses the data of the past, to understand the trend, and employing the data, it tries to make prediction based on various algorithms.

Machine learning can be broadly divided into two sub-categories namely: -

1.     Supervised learning

2.     Unsupervised learning

In supervised learning, the data, which we use for training our machine, is split into two parts. The first set, is used for training or "teaching" our model while the second set, is used for evaluating our trained model, by comparing the values generated by our model with the known values (often referred to as a ‘label’).

Whereas in the case of unsupervised learning, the whole of the data is used for training the model.

Also, an integral part of ML, are the algorithms one employs for training the machine.

Machine learning algorithms are broadly classified into the following three categories: -

1.     Classification: - This category simply deals with dividing the data into different categories based on a given criteria. A simple example of this could be, given the details of a patient, whether that patient is diabetic or not can be put under this category.

2.     Regression: - This algorithm deals with problems where a numerical output is expected. This deals with questions such as “how much” or “how many”.

3.     Clustering: - As the name suggests, clustering means putting the data into groups or clusters.

The backbone of machine learning is data analytics. Simply put, data analytics pertains to using the past data to draw results in the present, and also predict about the future.

Hence, it won’t be wrong to say that the data is of paramount importance to us. Using the data, the machine learns and subsequently uses, whatever it has learnt to come up with a result. 

P.S- This is my first article on LinkedIn. Constructive feedback and comments are welcome!