How Machines Learn—Explained Like You're 5

In this Knowledge nugget, we will explain few key AI concepts in a easy to digest manner.

Let's imagine you have a smart toy robot friend named Robi. You want to teach Robi new things so it can help you with fun tasks.

Learning with Help (Supervised Learning):

For example, you show Robi pictures of apples and bananas. Every time you show a picture, you tell Robi, "This is an apple" or "This is a banana." Robi remembers your words and the pictures. So, next time when Robi sees an apple or a banana, it can tell you which one it is! That's essentially how supervised learning works.

Learning by Finding Patterns (Unsupervised Learning):

In this scenario, you give Robi a box of colorful blocks but don't say anything. So, Robi looks at the blocks and notices that some are red, some are blue, and some are yellow. Robi then sorts the blocks into groups by color all by itself! This learning behaviour of the Robi is called unsupervised learning.

Learning from Trying (Reinforcement Learning):

Let's imagine you are playing a game where Robi needs to find the way through a maze to reach a yummy treat. And every time Robi makes a good move, you clap and say, "Good job!" If Robi hits a wall, you gently say, "Try again!" . Consiquently, Robi learns the best path to the treat by remembering what made you happy. This is the foundational principle of reinforcement learning.

AI is a huge field, there are more deeper concepts , methods and strategies being implemented now. However, the above three provides a high level overview of major concepts in AI. In the next week of Knowledge Nuggets, we will delve deeper and introduce few more interesting concepts from the world of AI. Stay tuned and subscribe to the Knowledge Nuggets newsletter if haven't already done so!