---
title: "Conformalized Rate-Adaptive Sensing: Ensuring Reconstruction Error Below Target with High Probability"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2607.26887v1"
permalink: "https://feed.astralor.com/items/itm_fb632648009c9c4eb3dfddd7"
published: "2026-07-29T13:16:27.000Z"
score: 76
channel: "finance"
tags: ["CoRAS","Compressed Sensing","Image Reconstruction","Imaging Systems","Conformal Prediction","Adaptive Sensing"]
---

# Conformalized Rate-Adaptive Sensing: Ensuring Reconstruction Error Below Target with High Probability

## AI commentary

> This paper cleverly applies conformal prediction to the adaptive stopping problem in imaging systems, balancing statistical rigor and practical utility. It offers a directly deployable framework for those working on medical imaging, remote sensing, or any resource-constrained compressed sensing. Notably, its calibration using early reconstruction behavior resembles an 'early exit' but with more reliability.

## At a glance

This paper proposes Conformalized Rate-Adaptive Sensing (CoRAS), a method to address the fundamental question in high-resolution imaging systems: when have enough measurements been collected. CoRAS adaptively chooses an acquisition/compression rate for each image using conformal prediction, keeping reconstruction error below a target level with high probability. It calibrates stopping time upper bounds using early reconstruction behavior, and experiments show target coverage, fewer average measurements, and adaptive allocation to harder images.

## Key points

- CoRAS adaptively determines acquisition rate per image via conformal prediction, ensuring reconstruction error below target.
- Estimates stopping time using early reconstruction path and calibrates with similar images for coverage guarantees.
- Achieves target coverage, fewer average measurements than fixed-rate rules, and allocates more measurements to harder images.

## Full text

Many high-resolution imaging systems face the same fundamental question: when have enough measurements been collected to reconstruct an image accurately? We develop Conformalized Rate-Adaptive Sensing (CoRAS), a method that adaptively chooses an acquisition or compression rate for each image while keeping the reconstruction error below a target level with high probability. As measurements are collected, an image reconstruction model gradually recovers the true image, producing a reconstruction path over acquisition rates. CoRAS uses this path up to an early decision time to estimate the target stopping time, defined as the first time at which the reconstruction error falls below the target level. It then calibrates this estimate using images with similar early reconstruction behavior, producing an upper bound on the stopping time with marginal and approximate conditional coverage guarantees. Experiments on image datasets show that CoRAS attains the target stopping-time coverage, uses fewer measurements on average than fixed-rate stopping rules, and assigns more measurements to images that are harder to reconstruct.

---

Verify against the original before citing; curated by Astral Feed: https://feed.astralor.com/items/itm_fb632648009c9c4eb3dfddd7