valeman/awesome-conformal-prediction

The collection organizes resources across multiple programming ecosystems—Python, R, and Julia libraries—alongside academic papers, theses, and conference materials spanning over a decade of research rooted in Kolmogorov's foundational work on algorithmic randomness. It emphasizes distribution-free uncertainty quantification methods that generate prediction intervals without parametric assumptions, making conformal prediction applicable across regression, classification, and time-series tasks. The curation draws from established venues like ICML, NeurIPS, and the dedicated COPA conference, positioning conformal prediction as a practical framework for reliable AI in production systems requiring formal coverage guarantees.

1,181 stars.