# PAC Confidence Sets for Deep Neural Networks via Calibrated Prediction

Source: [PAC Confidence Sets for Deep Neural Networks via Calibrated Prediction](<https://www.sota2.com/research/paper/68fa67041e209131cea13cd8>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T22:04:19.938Z

Authors: Sangdon Park, Osbert Bastani, Nikolai Matni, Insup Lee

Published: 2019-12-31T23:02:01.000Z

## Abstract

We propose an algorithm combining calibrated prediction and generalization bounds from learning theory to construct confidence sets for deep neural networks with PAC guarantees---i.e., the confidence set for a given input contains the true label with high probability. We demonstrate how our approach can be used to construct PAC confidence sets on ResNet for ImageNet, a visual object tracking model, and a dynamics model for the half-cheetah reinforcement learning problem.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2001.00106>)

[PDF](<https://arxiv.org/pdf/2001.00106.pdf>)