Improving Uncertainty Quantification of Deep Classifiers via Neighborhood Conformal Prediction: Novel Algorithm and Theoretical Analysis

TL;DR

This paper introduces Neighborhood Conformal Prediction (NCP), a new algorithm that enhances uncertainty quantification in deep classifiers by reducing prediction set size through adaptive, representation-based calibration, backed by theoretical guarantees and extensive experiments.

Contribution

The paper proposes NCP, a novel conformal prediction method that leverages neural network representations to produce smaller, more efficient uncertainty sets with theoretical validation.

Findings

NCP significantly reduces prediction set size compared to traditional CP methods.

Theoretical analysis shows NCP's prediction sets are smaller under mild conditions.

Experiments on CIFAR-10, CIFAR-100, and ImageNet confirm improved efficiency of NCP.

Abstract

Safe deployment of deep neural networks in high-stake real-world applications requires theoretically sound uncertainty quantification. Conformal prediction (CP) is a principled framework for uncertainty quantification of deep models in the form of prediction set for classification tasks with a user-specified coverage (i.e., true class label is contained with high probability). This paper proposes a novel algorithm referred to as Neighborhood Conformal Prediction (NCP) to improve the efficiency of uncertainty quantification from CP for deep classifiers (i.e., reduce prediction set size). The key idea behind NCP is to use the learned representation of the neural network to identify k nearest-neighbors calibration examples for a given testing input and assign them importance weights proportional to their distance to create adaptive prediction sets. We theoretically show that if the learned data representation of the neural network satisfies some mild conditions, NCP will produce smaller prediction sets than traditional CP algorithms. Our comprehensive experiments on CIFAR-10, CIFAR-100, and ImageNet datasets using diverse deep neural networks strongly demonstrate that NCP leads to significant reduction in prediction set size over prior CP methods.