SELF-SUPERVISED CONFORMAL PREDICTION WITH EQUIVARIANT BOOTSTRAPPING FOR IMAGE UNCERTAINTY QUANTIFICATION

Henry J. Aldridge, Tobías I. Liaudat, Marcelo Pereyra, Jason D. McEwen · 2026

Read SELF-SUPERVISED CONFORMAL PREDICTION WITH EQUIVARIANT BOOTSTRAPPING FOR IMAGE UNCERTAINTY QUANTIFICATION by Henry J. Aldridge, Tobías I. Liaudat, Marcelo Pereyra, Jason D. McEwen (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-18655.txt.

Description

Henry Aldridge and three co-authors introduce a 2026 method for gauging uncertainty in rebuilt images without ground-truth data for calibration. Inverse problems recover a signal from noisy, transformed observations, and many plausible images can fit the same data. Their approach builds rough coverages by using data symmetries through equivariant bootstrapping, then refines them with conformal calibration. Tested on weak-lensing mass mapping, it rebuilds the convergence field from galaxy shear measurements. The self-taught route avoids costly simulations tied to specific cosmological models.

Catalog record

CREATORHenry J. Aldridge, Tobías I. Liaudat, Marcelo Pereyra, Jason D. McEwen
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30129
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.18655

Browse the full library · About the Tower of Babel