Self-Conditioned Denoising for Atomistic Representation Learning

Tynan J. Perez, Rafael Gómez-Bombarelli · 2026

Read Self-Conditioned Denoising for Atomistic Representation Learning by Tynan J. Perez, Rafael Gómez-Bombarelli (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-2603-17196.txt.

Description

Perez and Gomez-Bombarelli present Self-Conditioned Denoising, a pretraining method for atomistic models. Pretraining has changed language and vision, but atomistic models still lean on costly labeled force and energy data. Their method guides denoising on the model's own embeddings. It works across domains, from small molecules and proteins to periodic solids. With backbone and data held fixed, it beat prior self supervised methods on downstream benchmarks. It matched or beat supervised force energy pretraining.

Catalog record

CREATORTynan J. Perez, Rafael Gómez-Bombarelli
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE63291
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.17196

Browse the full library · About the Tower of Babel