AI engineering glossary · ML fundamentals
What is true risk?
True risk (also called Expected risk, Population risk): The expected loss a model would incur over the whole distribution of data it could ever meet, not just the sample it was trained on. It is the quantity we really care about but can never compute exactly; test data gives an estimate of it.
True risk: an example
A spam filter's true risk is its average error over every email that will ever arrive, whereas 2,000 test emails give only an estimate.
Related terms
More ML fundamentals terms
- Underfitting
- Unsupervised learning
- Validation curve
- Validation set
- VC dimension
- Weak supervision
- Ablation study
- Active learning
Where you learn it
True risk is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.