PAC Confidence Predictions for Deep Neural Network Classifiers

TL;DR

This paper introduces a new method for providing reliable confidence estimates for deep neural network predictions, with formal guarantees, enabling safer and more efficient deployment in critical applications.

Contribution

The paper presents a novel algorithm combining Clopper-Pearson intervals and histogram binning to produce calibrated confidence predictions with provable correctness guarantees.

Findings

Provides guarantees for state-of-the-art DNNs' confidence estimates

Enables composition of fast and accurate DNNs for improved inference

Guarantees safety in DNN-based decision making in visual tasks

Abstract

A key challenge for deploying deep neural networks (DNNs) in safety critical settings is the need to provide rigorous ways to quantify their uncertainty. In this paper, we propose a novel algorithm for constructing predicted classification confidences for DNNs that comes with provable correctness guarantees. Our approach uses Clopper-Pearson confidence intervals for the Binomial distribution in conjunction with the histogram binning approach to calibrated prediction. In addition, we demonstrate how our predicted confidences can be used to enable downstream guarantees in two settings: (i) fast DNN inference, where we demonstrate how to compose a fast but inaccurate DNN with an accurate but slow DNN in a rigorous way to improve performance without sacrificing accuracy, and (ii) safe planning, where we guarantee safety when using a DNN to predict whether a given action is safe based on visual observations. In our experiments, we demonstrate that our approach can be used to provide guarantees for state-of-the-art DNNs.