Human beings learn from Active Experimentation, Concrete Experience, Abstract Conceptualization and Reflective Observation. Based on these four factors, further a human being learns to recognize patterns in various phenomenon around them.

Human beings replicate their experience or use discretion amidst variable circumstances. So, if I have to hit a moving target there are certain variables about which I have previous learning and I use this example and take a decision.

There are visual and kinaesthetic components involved in making decisions. My decision of hitting a moving target will depend on the speed of the object, the air resistance, the rotation of the object and the revolution of the object. So, there is a predictive decision that is made and based on that an action is taken.

Human beings can understand a finite number of variables with limited accuracy.

What if we have infinite variables and are in need of very high accuracy?

We need to resort to 2 things:

  1. Standardize certain variables and work around with the others
  2. Improve the Human Potential to challenge the variables

Both these things are difficult, but the challenge of variables is growing by the day.

The best thing can be to use a tool that replicates human behaviour and at the same time extrapolate the potential to handle infinite variables.

Here we arrive at Machine Learning.

Traditionally, we used machines for transaction processing, calculations, and computations, then we moved to standardizing activities. The next generation of machines is making decisions available where it is to possible take decisions based on. A calculator is the first known electronic machine which standardized human effort in computation, so the human effort was involved in completing tasks.

With the advent of an information revolution, tasks were standardized, so the human effort was put into making decisions. Now with the advent of Industry 4.0, the effort to arrive at decisions is replaced with Artificial Intelligence (AI). Moving further, to eliminate the entire process of effort, task and decision standardization, the generation has moved to standardize Learning so that human effort is eliminated among complex data analysis and decision making.

Welcome to the world of standardizing knowledge and learning – we call it the Machine Learning. Thus, Machine Learning is nothing but it is the subfield of Artificial Intelligence.

Human Brain is a hard disk on which data is written based on Active Experimentation, Concrete Experience, Reflective Observation and Abstract Conceptualization.

  • Active Experimentation is a Kinaesthetic Technique of Learning
  • Concrete Experience is a learning based on repetitive practice
  • Reflective Observation is a method of Extrapolating and Interpolating
  • Abstract Conceptualization is a method of predictive learning

The objective of Machine Learning is to make Machines replicate this style of learning and thereby use the relevant behaviours in various situations. The Machine learns to behave based on Experience it gains by recognizing patterns. It is an approach to AI that’s focused on making machines which can learn without being explicitly programmed.

Types of Machine Learning

There are three different types of Machine Learning algorithms, namely Supervised Learning, Unsupervised Learning and Reinforcement Learning.

  1. Supervised Learning – The system is provided with examples in the form of inputs with a desirable outcome. Here the system tries to learn from the previous examples, compare output, find an error and modify the prototype accordingly. The common application of supervised learning includes the prediction of future events like a weather report, stock market fluctuations or as simple as filtering of spam emails.
  2. Unsupervised Learning – Here the system is not provided with any outcome or right answer and it is left on its own to structure the data and find the answer. The algorithm is left to find commonalities among the given complex data or input to structure in a meaningful manner. For example, an unsupervised learning algorithm can be used to detect fraudulent credit card transactions.
  3. Reinforcement Learning – In reinforcement learning the system interacts with the dynamic environment and perform certain tasks where it receives input in the form of feedback and finds a solution using trial and error method, like playing a game or driving a vehicle.


Machine Learning is one of the sought-after application of AI. We rely on this technology in our day to day tasks such as traffic predictions, virtual assistance, social media services, online customer support etc… Machine Learning is continuously innovating and evolving and its algorithms, methods, and approaches change with time.

In this blog, we covered an introduction to Machine Learning and three common types of machine learning. You can learn more about its common algorithms, approaches, programming languages in machine learning and applications in the coming weeks.

      Stay tuned for more !!

     If you have enjoyed reading our blog, do leave your comments below.