What is Machine Learning and How does it works?
Table of Contents
We are living in the era of humans and machines. At one end human have been learning from their past experiences since millions of years on the other hand era of machines has just begun. Today these machines need to be preprogrammed in order for them to carry out tasks or follow instructions. But what if we could have machines that learn and perform on their own, this is where machine learning comes in.
What is Machine Learning?
Machine Learning is a field based on research, devoted to build methods that can understand and learn. It is branch that works to build methods that can use data to improve performance on set of tasks. Machine learning algorithms are utilized in a broad range of applications, including medicine, email filtering, speech recognition, and computer vision, where developing traditional algorithms to do the required tasks is difficult or impossible. There are various Machine Learning tools which make it easy for us to build such algorithms.
How does Machine Learning works?
The figure below briefly explains how machine learning works. Generally, a labelled or unlabeled data is modelled so that it can be used effectively in Machine Learning Algorithm. This modelled data is then sent into Algorithm which will give some kind of prediction as output. This prediction is then evaluated for accuracy and if accuracy is acceptable the machine learning algorithm is then deployed. If the accuracy is unacceptable then machine learning algorithm is strained again and again with a training data set until it generates desired prediction accurately.
AI vs ML vs Deep Learning
Many people think that all the three concepts Artificial Intelligence, Machine Learning and Deep Learning are the same. But they are wrong AI is much broader concepts compared to others. Machine Learning here means to allow machines to identify meaningful patterns from data and learn from that. Whereas Artificial Intelligence is about giving human intelligence to machines. This could also mean that all Machine Learning is AI but not all AI is Machine Learning. Deep Learning here means applying similar algorithm on Neural Networks to get better accuracy compared to simple ML applications. Deep Learning is a subset of Machine Learning. It is generally based on neural networks which tend to work like neurons of human brain thus producing prediction with best accuracy.
Some of the examples of AI from our day-to-day life are Apple Siri, Chess playing computer, Tesla’s self-driving car etc. While some of the examples of Machine Learning are Snapchat’s filters for your flower crown selfies based on augmented reality and Netflix’s recommendation system that works based on your past data. Let us also understand deep learning with an example: Suppose you are told to identify whether a particular figure is a square or not, what would you do? First of all, you would check whether the figure has four sides, then you might check whether all four sides are equal, going further you can check if there is a right angle between adjacent sides. Here based on above checks you can build a model which can take multiple figures and predict whether particular figure is square or not. But what if you would even not have to give these checks to model and model itself will figure out what a square figure should have and what not, that particular type of model is said to be Deep Learning Model. Deep Learning is generally based on neural networks which tend to work like neurons of human brain thus producing prediction with best accuracy.