Brief Introduction to the term “Machine Learning”
The term Machine Learning has a deep correlation with Artificial Intelligence. A vast majority of you may have the intuition that ‘they are in fact the same’. But they are not. The term Artificial Intelligence refers to the intrusion of intelligence on a machine which allows it to simulate human-like behaviour. Machine Learning is considered as a subset of Artificial Intelligence. In Machine Learning, we are basically training a machine to exhibit intelligence. Like the term suggests, the machine learns the task rather than allowing developers to program explicitly. Since the whole process is about learning, it needs some past data to model the task. This in short gives you an abstract idea about Machine Learning.
Before we move further, I would like to point out a few basic intuitions regarding our topic. We often have a commotion whether Machine Learning solely belongs to the Computer Science branch or not. You should first obliterate this thought as Machine Learning is an interdisciplinary topic. Most importantly, I would firmly shout out that Mathematics is the very soul of Machine Learning. The latter is just a tool to put it into reality. A strong mathematical understanding is imperative in-order to understand what really happens inside Machine Learning domain. You may also have the option to develop Machine Learning models even without any prior knowledge on the basic mathematics related to it. But trust me, you will miss the very identity of the domain and rather end up in the conclusion that this domain is nothing short of a normal day to day coding domain, which Machine Learning is not.
Machine Learning revolves around modelling a particular task using a related predefined data and further using this model to make informed decisions. Have you guys ever wondered how you are capable of categorising different animals? Let us just take a simple example of a cat and a dog. You can clearly identify and categorise them without much effort. It is solely because you are familiar with the key features that distinguish a cat and a dog, either from your real life experience or from some past experiences. Bingo! This in fact is the basic idea behind Machine Learning. Train from an existing data, develop a model and do the inference.
It is crystal clear that a predefined data is used for modelling the tasks. What if you are taught with something which is totally irrelevant for your task. The same scenario can be extended to the level of Machine Learning as well. In our case, the nature of the data which we rely on training the model plays a key role along with the mathematical model upon which we train. Collecting the data alone is not sufficient, you should always ensure that it contains the utmost relevant information about your task. A fair and a deep study on the type and nature of the available data is required before directly stepping on to the modelling part. Once the data handling part is completed, you can build the model using these verified data. The data handling part is collectively called as data pre-processing. With these, I welcome all the beginners to the world of Machine Learning.
In this blog, I have presented you with the basic idea of the term “Machine Learning” and I hope this blog was helpful and would have motivated you enough to get interested in the topic.
