Cross-conformal predictors

By Vladimir G Vovk. 2012

262 citations recorded by the source corpus.

Inductive conformal predictors have been designed to overcome the computational inefficiency exhibited by conformal predictors for many underlying prediction algorithms. Whereas computationally efficient, inductive conformal predictors sacrifice different parts of the training set at different stages of prediction, which affects their informational efficiency. This paper introduces the method of cross-conformal prediction, which is a hybrid of the methods of inductive conformal prediction and cross-validation, and studies its validity and informational efficiency empirically. The computational efficiency of cross-conformal predictors is comparable to that of inductive conformal predictors, and they produce valid predictions in our empirical studies.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

Citations in the corpus

DeltaGrad: Rapid retraining of machine learning models (2020, methodology): “As an example, in cross-conformal prediction (Vovk, 2015) we have a predictive model f ˆ that can be trained on any subset of the data.”
Predictive inference with the jackknife+ (2019, methodology): “Our proposed CV+ prediction interval is related to the cross-conformal prediction method of Vovk [27], Vovk et al.”
A Gentle Introduction to Conformal Prediction and Distribution-Free Uncertainty Quantification (2021, background): “However, the guarantee is more brittle than that of conformal prediction; see [55] and [61].”
Combining p-values via averaging (2012, methodology): “ p1,...,pK.) Intuitively, the most natural way to combine K numbers is simply to average them; essentially, this is the way of combining p-values used in the method of cross-conformal prediction (see [14], (11)). None of the functions F in (1), (2), and (3) involves the average ¯p := 1 K (p1+···+pK)”
Conformal prediction beyond exchangeability (2022, background): “1 Background: proof of Theorem 1 To help build intuition for the proof techniques we will use later on, we reformulate Vovk et al. [2005]’s proofs of these results, and we then explain some of the challenges in extending these existing results to our new setting.”
Conformal inference of counterfactuals and individual treatment effects (2020, methodology): “Conformal inference was introduced by Vladimir Vovk and his collaborators (e.g. Balasubramanian et al., 2014; Gammerman & Vovk, 2007; Shafer & Vovk, 2008; Vovk, 2012; Vovk, 2013; Vovk, 2015; Vovk et al., 2009).”