Machine Learning (ML)
Teaching a computer by showing it lots of examples instead of writing step-by-step instructions.
In everyday terms
Rather than coding "spam contains these words", you show the system thousands of spam and non-spam emails and it works out the pattern itself.
For professionals
Algorithms that optimise a model's parameters to minimise error on training data, aiming to generalise to unseen data. Main flavours: supervised, unsupervised and reinforcement learning.
Think of it like…
Learning to recognise dogs by seeing hundreds of dogs, not by reading a definition of "dog".
You've already seen it
Bank fraud alerts, "customers also bought", voice assistants learning your accent.
Myth vs reality
Myth: Machine learning means the computer is programmed with knowledge.
Reality: It is programmed with a way to learn; the knowledge comes from data.
Quick check
What does a machine-learning system mainly learn from?
- Rules typed by programmers
- Examples (data)
- The internet in real time
- Its own feelings
Show answer
Examples (data): ML extracts patterns from example data.