A 3 B 2 : Adaptive Asymmetric Adapter for Alleviating Branch Bias in Vision-Language Image Classification with Few-Shot Learning
Yiyun Zhou, Zhonghua Jiang, Wenkang Han, Kunxi Li, Mingjing Xu, Chang Yao, Jingyuan Chen · 2026
Read A 3 B 2 : Adaptive Asymmetric Adapter for Alleviating Branch Bias in Vision-Language Image Classification with Few-Shot Learning by Yiyun Zhou, Zhonghua Jiang, Wenkang Han, Kunxi Li, Mingjing Xu, Chang Yao, Jingyuan Chen (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-2605-13161.txt.
Description
Zhou and six co-authors address Branch Bias in few-shot vision-language image classification: tuning the image encoder does not always help when data shifts. Their adapter, A3B2, damps image-branch tuning when prediction uncertainty is high, with no manual intervention. The design is light and asymmetric, inspired by mixture-of-experts with load balancing. Across three few-shot tasks on 11 datasets it consistently beats 11 prompt- and adapter-based baselines.
Catalog record
| CREATOR | Yiyun Zhou, Zhonghua Jiang, Wenkang Han, Kunxi Li, Mingjing Xu, Chang Yao, Jingyuan Chen |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68911 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.13161