Hey there! Welcome back to another episode of Tech Break by Friday. Today, we have another episode of the series AI Demystified. Our topic: What’s the real difference between Artificial Intelligence, Machine Learning, and Reinforcement Learning? Spoiler: they’re not all the same thing, and understanding the difference will help you spot the hype from the real deal.
(Source: Created with ChatGPT)
Let’s remember what AI means from our last episode. AI is an algorithm that recognises patterns based on data that it has already seen. The period that AI learns from data is called the training period. After training, AI is capable of recognising the patterns in the data. An example that you might already have seen is the Netflix recommendation system. Based on the movies you have seen, it proposes new films and series that you might like.
🧠 What is Machine Learning (ML)?
Machine Learning (ML) is a part of artificial intelligence (AI) that allows computers to learn patterns from data and make decisions without being explicitly programmed for each task. Unlike traditional software, which follows instructions (“if A, then B”), (“else C”).
A widely cited definition by Tom Mitchell states: “A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P if its performance at tasks in T, as measured by P, improves with experience E.” For example, an email spam filter is an ML solution. The spam email filter is the task (T). The ML algorithm gets better at recognising spam emails ( this is the performance (P)). Performance here means how many emails have been recognised correctly. The data that has already been seen or processed, labelled emails of “spam”, “not spam”, are the experience (E).
Core Types of Machine Learning (ML)
Supervised Learning
In supervised learning, models are trained using labelled data—each input has a known output. The algorithm learns to map inputs to outputs by minimizing prediction errors. Typical applications include:
Classification: Sorting emails as spam or non-spam.
Regression: Predicting house prices based on features like size and location. (The difference here is that the data are not countable like spam, non-spam)
The model uses its training to make predictions on new, unseen data.
Unsupervised Learning
Unsupervised learning deals with unlabeled data. The algorithm tries to find hidden patterns or groupings within the data. Standard techniques include clustering (grouping similar customers). This approach is often used for anomaly detection. For example, you are given different shapes of objects manufactured. The unsupervised learning method explores all possible options to recognise which of these objects were created with anomalies.
Reinforcement Learning
Reinforcement learning (RL) is about learning by doing. An agent interacts with an environment, making decisions and receiving feedback in the form of rewards or penalties. Over time, the agent learns which actions yield the best outcomes. RL is used in robotics, game-playing AIs like AlphaGo, and industrial automation, where trial and error can optimise performance.
Think of a video game where you start with no instructions — just a controller and a goal to survive or score points.
Every time your character jumps over a pit and survives, you earn points — great! But when you walk into a trap, you lose a life — not so great.
Over time, you learn what actions help you win: jump at the right moment, avoid certain areas, grab bonus items.
That’s reinforcement learning:
The game is the environment
You are the agent
Points and penalties are your feedback (rewards)
🎯 In real life: This is how AIs like AlphaGo learn to master complex games, or how robots learn to walk by trying, failing, and gradually improving with each attempt.
If you found it helpful and you liked this episode, share it with a friend. It helps more people find the podcast and learn about AI. Subscribe, leave a review, and keep building.
Sources:
SAP: What is Machine Learning?: https://www.sap.com/germany/products/artificial-intelligence/what-is-machine-learning.html
Dev with Darwish – Core Machine Learning Paradigms (YouTube, April 2025)
OPIT: Supervised vs. Unsupervised Learning https://www.opit.com/magazine/supervised-vs-unsupervised-learning/
Alexander Thamm: Unsupervised Learning Explained https://www.alexanderthamm.com/en/blog/this-is-how-unsupervised-machine-learning-works/
Bernard Marr: What Is Reinforcement Learning https://bernardmarr.com/artificial-intelligence-what-is-reinforcement-learning-a-simple-explanation-practical-examples/
Spiceworks: What is machine learning?https://www.spiceworks.com/tech/artificial-intelligence/articles/what-is-ml/
Datasolut: Machine Learning Definition https://datasolut.com/was-ist-machine-learning/
Share this post