Episode 15: AI Demystified #2 - What is AI vs. ML vs. Reinforcement Learning?

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)