Split Localized Conformal Prediction
This work addresses a key limitation in uncertainty quantification for machine learning practitioners, though it represents an incremental improvement over existing split conformal methods.
The paper tackles the problem of achieving conditional coverage guarantees in conformal prediction, which is stronger than average coverage but typically requires complex additional models. The proposed method uses kernel density estimation to modify non-conformity scores, resulting in improved performance in both average and conditional coverage metrics.
Conformal prediction is a simple and powerful tool that can quantify uncertainty without any distributional assumptions. Many existing methods only address the average coverage guarantee, which is not ideal compared to the stronger conditional coverage guarantee. Existing methods of approximating conditional coverage require additional models or time effort, which makes them not easy to scale. In this paper, we propose a modified non-conformity score by leveraging the local approximation of the conditional distribution using kernel density estimation. The modified score inherits the spirit of split conformal methods, which is simple and efficient and can scale to high dimensional settings. We also proposed a unified framework that brings together our method and several state-of-the-art. We perform extensive empirical evaluations: results measured by both average and conditional coverage confirm the advantage of our method.