Hyperparameters
Settings chosen before training that shape how a model learns.
Examples are learning rate, number of layers and tree depth. Different settings produce different models, so they should be documented and versioned for reproducibility.
On the exam
Undocumented hyperparameters make a model impossible to reproduce or re-validate.
Related terms
- Model registry and versioningRecording each model version with its data, code, settings and approvals.
- Overfitting and underfittingA model too tailored to its training data, or too simple to capture the pattern.
AAIA Test yourself on this with timed AAIA practice questions in the AAIA Exam Simulation Hub, or read the free AAIA exam guide.
Test yourself with today’s question of the day.
