Conformal Prediction: The Definitive Resource
A curated list of conformal prediction resources: tutorials, videos, articles, libraries, courses, books and papers on this distribution-free framework for uncertainty quantification in machine learning.
This list collects resources on conformal prediction, a distribution-free, model-agnostic framework for quantifying predictive uncertainty. Instead of a single point prediction, a conformal predictor returns a prediction set or interval that is guaranteed to contain the true label with a chosen probability, such as ninety or ninety-five percent, under the assumption that the data are exchangeable, and this guarantee holds in finite samples for any underlying model without distributional assumptions.
It is organized as a learning path: a short conceptual introduction, a recorded talk, a gentle written introduction, library quickstarts in Python, and finally courses, books and papers for going deeper. Topics referenced across the collected material include prediction intervals, model calibration, selective prediction, time series forecasting, anomaly detection and uncertainty in large language models.
The list also links to a paid cohort-based course and a practitioner's book by one of the field's researchers.