top of page

Hands on machine learning

Introduction : Machine learning is an exciting branch of Artificial Intelligence, and it’s all around us. Machine learning brings out the power of data in new ways, such as Facebook suggesting articles in your feed. Machine learning is the field to study that gives computers the ability to learn without being explicitly programmed. As you input more data into a machine, this helps the algorithms teach the computer, thus improving the delivered results. When you ask Alexa to play your favourite music station on Amazon Echo, she will go to the station you played the most.

Types of Machine Learning techniques

Whether or not they are trained with human supervision, machine learning models are divided into three parts -

Supervised Learning: Supervised learning, as the name suggests, has the presence of a supervisor. Basically, in this machine learning technique the model is fed data and its labels for training.

Unsupervised Learning: Unsupervised learning is the technique where we don’t specify the labels for the training data. In this method, a model tries to learn to predict or classify on its own.

Semi-Supervised Learning: This algorithms are trained with partially labelled data, so some of them are labelled and some of them are not.


Learning: Reinforcement learning is something else. It completely learns through the results. It observes, then selects an action using a policy and then takes action. It gets a reward or penalty. It updates the policy(this is the learning step of it). It iterates until an optimal policy is found.

Reinforcement Learning:

Another area of machine learning techniques is reinforcement learning. Here unlike supervised learning, it doesn’t have a label for the data. The model decides itself what to do and based on the results it sends back a feedback to make changes for the best result. In the absence of a training dataset, it is bound to learn from its experience.

In this machine learning technique an agent learns to behave in an environment by performing actions and customising the results output. The events are classified as rewards that means the agent is learning right things, and punishment which is more like a feedback to the agent to change their behaviour.

Two main components :

Agent - The Reinforcement Learning algorithm that learns from trial and error.

Environment - The world through which the agent moves.

Agent Training Theories

There are different theories about training the agent. Reward Maximization theory states that a agent must be trained in a way that it takes the best action so that the rewards are maximized.

Another theory is exploration, which is about exploring and capturing information about the environment.

Use cases of Reinforcement Learning:

This is used in robotics for industrial automation.

Parking a car

Driving tasks

This method is also used in data processing.

Dynamic Pathing

Controller optimization


The machine learning concepts seems pretty hard at first but as we dive deeper into the concept and apply it on our own this becomes easy. Creating machine learning models are more like trial and error, we have different methods to get best results and best accuracy. We have different functions to test the model on test data and validation data. If someone thinks it’s all about math he might be right but, to apply these concepts you don’t need mathematics. To know how the models actually works from inside and to learn the core concepts of these models you need knowledge on mathematics for sure.

Neural Networks are a different topic that’s beyond the scope of this article. This is a whole another level of concepts and ideas, to start with Neural Networks or Computer Vision ideas do checkout the articles here -

25 views0 comments

Recent Posts

See All


bottom of page