Machine Learning as a subset of artificial intelligence

AI (artificial intelligence) and ML (machine learning) are two very trendy words these days and often appear to be used interchangeably. However, they are not the same. Artificial intelligence is the broader notion of the machines accomplishing tasks “intelligently” whereas machine learning is based on the intention that we should just give machines the access to data and let them comprehend it for themselves.  But what is machine learning exactly? Let’s clarify this once and for all!

Machine learning is one of the most riveting technologies that we have come across until now and it evolved as a subset of AI from pattern recognition technology. As humans, we take pride in having the “ability to learn”, but as of now, that ability is also being passed on to our non-living next of kin.

The term ML was assigned in 1959 by Arthur Samuel, who was an American forerunner in the field of gaming and AI. From Alan Turing’s proposition in his published paper, we conclude that the question has now changed from “can machines think?” to “can a machine do what we (as thinking entities) can do?”

It can be commonly categorized into four categories:

I. Supervised learning

II. Unsupervised learning

III. Reinforcement learning


IV. Semi-supervised learning

Machine Learning courses

Supervised learning: A machine learns from a specimen data and its associated target answer to be able to foresee a correct solution when it is presented with new examples. This type of learning comes under supervised learning.

Unsupervised learning: A machine can also learn from being given a specimen data and left to interpret its pattern and find meaning on its own. This type of learning comes under unsupervised learning. This is rather beneficial in providing us with meaningful insights into the data.

Reinforcement learning: This generally involves a machine learning through unsupervised learning methods and then is given positive or negative feedback based on the solution the machine infers. A very simple example of this type of learning can be computer learning to play the video game.

Semi-supervised learning: In this, the machine is given an incomplete signal along with sets of target responses that are missing some parts of it as well.

The expected role of ML in our lives:

  • Speech recognition (Cortana, Siri etc. make use of what’s called “Natural Language Processing” or “N.L.P.”)
  • Computer vision (machine’s probable understanding of our real world such as facial recognition, pattern recognition, etc.)
  • Google’s self-driving car



 Now, for the unexpected:


  • Recommendations (Amazon, YouTube, Netflix, etc. all make use of the learning algorithm called “Recommender Systems”. It basically reads into each and every user’s personalized preference and makes suggestions based on that.)
  • Stock market/housing finance/real estate (all of these integrate ML into their systems to better anticipate the market situation with what’s called “Regression Techniques”.)


 Of course, machine learning has shortcomings and the technology is far from succeeding with the achievement of true AI. But despite this, ML continues to be an effective way to turn data into valuable understandings. From research & technology to improving businesses, machine learning courses is useful everywhere. And henceforth, it also is quite a good option as a profession. The industry is on the rise and does not seem to be stopping anytime soon.


ExcelR – Data Science, Data Analytics Course Training in Bangalore

               49, 1st Cross, 27th Main,

               Behind Tata Motors, 1st Stage,

               BTM Layout, Bengaluru,

               Karnataka 560068

               Phone: 096321 56744

               Hours: Sunday – Saturday 7AM – 11PM

Leave a Reply

Your email address will not be published. Required fields are marked *