Hierarchy-Guided Multimodal Representation Learning for Taxonomic Inference

Sk Miraj Ahmed, Xi Yu, Yunqi Li, Yuewei Lin, Wei Xu · 2026

Read Hierarchy-Guided Multimodal Representation Learning for Taxonomic Inference by Sk Miraj Ahmed, Xi Yu, Yunqi Li, Yuewei Lin, Wei Xu (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-2603-25573.txt.

Description

Ahmed and co-authors teach species naming models to respect the tree of life. Field ID mixes specimen photos and DNA barcodes, but flat label training ignores how order, family, and genus nest. One variant shapes the embedding space level by level, so near kin stay near. A second adds a light fusion head that works with photo, DNA, or both. Across large biodiversity sets, accuracy beats strong rivals by over fourteen points. Gains grow when the DNA input is partial or corrupted.

Catalog record

CREATORSk Miraj Ahmed, Xi Yu, Yunqi Li, Yuewei Lin, Wei Xu
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE34293
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.25573

Browse the full library · About the Tower of Babel