This repository is a comprehensive, professionally curated list of resources on Conformal Prediction, targeting machine learning researchers, engineers, and practitioners. It aims to serve as an all-in-one guide for mastering this uncertainty quantification technique, offering a vast collection of videos, tutorials, books, papers, theses, articles, and open-source libraries across Python, R, and Julia.

How It Works

Conformal Prediction is a framework for generating prediction intervals with guaranteed coverage, meaning that for a given confidence level (e.g., 95%), the true value will fall within the predicted interval 95% of the time, regardless of the underlying data distribution. This is achieved by quantifying the "nonconformity" of new data points relative to a calibration set, ensuring validity without distributional assumptions.

Quick Start & Requirements

This repository is a curated list, not a software package. Installation and usage depend on the specific libraries or tools referenced within the list. The primary requirement is an interest in Conformal Prediction and its applications.

Highlighted Details

• Extensive collection of resources, including over 1000 papers, 100+ theses, and numerous tutorials and videos.
• Covers a wide range of applications, from time series forecasting and anomaly detection to NLP and medical imaging.
• Features contributions and endorsements from leading researchers in the field, such as Vladimir Vovk, Emmanuel Candes, and Larry Wasserman.
• Includes links to active courses, books, and software libraries for practical learning and implementation.

Maintenance & Community

The repository is maintained by Valery Manokhin, a PhD candidate specializing in Conformal Prediction. The project encourages community contributions and engagement through social media platforms like LinkedIn and Twitter.

Licensing & Compatibility

The repository content is licensed under CC BY-NC-ND 4.0. This license permits non-commercial use and redistribution but requires attribution and prohibits derivative works.

Limitations & Caveats

The repository is a curated list and does not provide direct software functionality. Users must navigate to external resources for implementation. Some listed resources may be outdated or require specific technical environments.