DAMEL: Dual-Axis Multi-Expert Learning for Class-Imbalanced Learning Journal: Knowledge-Based Systems

Hyuck Lee, Taemin Park, Heeyoung Kim · 2026

Read DAMEL: Dual-Axis Multi-Expert Learning for Class-Imbalanced Learning Journal: Knowledge-Based Systems by Hyuck Lee, Taemin Park, Heeyoung Kim (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-2605-30135.txt.

Description

Hyuck Lee, Taemin Park and Heeyoung Kim propose DAMEL, a dual-axis multi-expert method for imbalanced, long-tailed data. Rebalancing tricks cut bias but often raise variance, and existing multi-expert fixes add complex procedures. DAMEL attacks both problems with experts along two axes. On the representation axis it concatenates expert representations and trains a balanced classifier alongside them. On the time axis it averages network weights across epochs for use at test time. Experiments show it reduces both bias and variance.

Catalog record

CREATORHyuck Lee, Taemin Park, Heeyoung Kim
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49810
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.30135

Browse the full library · About the Tower of Babel