Intermediate · Fundamentals
Calibration layer
Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.
TL;DR. An adjustment made after a model's predictions to correct for bias and align probabilities with observed outcomes.
Technical Definition
An adjustment made after a model's predictions to correct for bias and align probabilities with observed outcomes.
How it works
A calibration layer is applied after a model has made its initial predictions. Its purpose is to adjust these predictions, often to correct for systematic biases. This adjustment ensures that the model's reported probabilities more accurately reflect the true likelihood of different outcomes based on observed data.