Conformalized Credal Regions for Classification with Ambiguous Ground Truth

TL;DR

This paper introduces a conformal prediction method for constructing credal regions in classification tasks with ambiguous ground truth, providing guarantees and disentangling uncertainty sources.

Contribution

It extends classical conformal prediction to handle ambiguous labels, enabling the construction of credal regions with theoretical guarantees and improved prediction sets.

Findings

Provides conformal coverage guarantees

Achieves smaller prediction sets than classical methods

Disentangles epistemic and aleatoric uncertainty

Abstract

An open question in \emph{Imprecise Probabilistic Machine Learning} is how to empirically derive a credal region (i.e., a closed and convex family of probabilities on the output space) from the available data, without any prior knowledge or assumption. In classification problems, credal regions are a tool that is able to provide provable guarantees under realistic assumptions by characterizing the uncertainty about the distribution of the labels. Building on previous work, we show that credal regions can be directly constructed using conformal methods. This allows us to provide a novel extension of classical conformal prediction to problems with ambiguous ground truth, that is, when the exact labels for given inputs are not exactly known. The resulting construction enjoys desirable practical and theoretical properties: (i) conformal coverage guarantees, (ii) smaller prediction sets (compared to classical conformal prediction regions) and (iii) disentanglement of uncertainty sources (epistemic, aleatoric). We empirically verify our findings on both synthetic and real datasets.