AI glossary
Machine learning
A way of building AI where the system learns patterns from examples instead of following rules written by hand.
In machine learning you do not program the answer; you show the system many examples and let it adjust its internal parameters until its predictions match. The result is a model you can use on new data.
There are three classic flavours: supervised learning (examples come with the right answer), unsupervised learning (find structure without labels) and reinforcement learning (learn from rewards).
Example: A bank trains a model on past transactions labelled “fraud” or “not fraud”, then uses it to flag suspicious new ones.
In practice
- Data quality rules: a model also learns the mistakes and biases in its examples.
- Test the model on data it did not see during training; otherwise it will look better than it is.
- Models go stale when reality changes (new kinds of fraud, for example) and need retraining.
There are free courses to get started in resources.