OgBench: A Framework for Evaluating Graph Neural Networks on Omics Data

Louisa CornelisUC Santa Barbara [email protected] Johan MatheAtmo, Inc.Louis Van LangendonckUniversitat Politècnica de Catalunya, Guillermo BernárdezUC Santa Barbara, Nina MiolaneUC Santa Barbara · 2026

Read OgBench: A Framework for Evaluating Graph Neural Networks on Omics Data by Louisa CornelisUC Santa Barbara [email protected] Johan MatheAtmo, Inc.Louis Van LangendonckUniversitat Politècnica de Catalunya, Guillermo BernárdezUC Santa Barbara, Nina MiolaneUC Santa Barbara (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-2605-15511.txt.

Description

Louisa Cornelis, Johan Mathe, Louis Van Langendonck, Guillermo Bernárdez, and Nina Miolane launch OgBench, a test bed for graph nets on omics data. Omics graphs are huge while patient samples are few, the flip of the usual case. Their platform sets a standard path from raw omics data to featured graph families. It tests classic graph nets, big-graph variants, and plain rivals. Widely used graph nets often fail to beat plain multilayer perceptrons. Graph structure thus does not always add value here.

Catalog record

CREATORLouisa CornelisUC Santa Barbara [email protected] Johan MatheAtmo, Inc.Louis Van LangendonckUniversitat Politècnica de Catalunya, Guillermo BernárdezUC Santa Barbara, Nina MiolaneUC Santa Barbara
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE89359
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.15511

Browse the full library · About the Tower of Babel