A$_3$B$_2$: Adaptive Asymmetric Adapter for Alleviating Branch Bias in Vision-Language Image Classification with Few-Shot Learning

TL;DR

This paper introduces A3B2, an adaptive asymmetric adapter that reduces branch bias in vision-language models for few-shot image classification, improving performance across multiple datasets.

Contribution

The paper proposes A3B2, a novel adaptive adapter with uncertainty-aware dampening and asymmetric design to address branch bias in vision-language few-shot learning.

Findings

A3B2 outperforms 11 baselines on 11 datasets.

UAAD automatically suppresses image-branch adaptation under high uncertainty.

Extensive experiments validate the effectiveness of A3B2.

Abstract

Efficient transfer learning methods for large-scale vision-language models ($e.g.$, CLIP) enable strong few-shot transfer, yet existing adaptation methods follow a fixed fine-tuning paradigm that implicitly assumes a uniform importance of the image and text branches, which has not been systematically studied in image classification. Through extensive analysis, we reveal a Branch Bias issue in vision-language image classification: adapting the image encoder does not always improve performance under out-of-distribution settings. Motivated by this observation, we propose A$_3$B$_2$, an Adaptive Asymmetric Adapter that alleviates Branch Bias in few-shot learning. A$_3$B$_2$ introduces Uncertainty-Aware Adapter Dampening (UAAD), which automatically suppresses image-branch adaptation when prediction uncertainty is high, enabling soft and data-driven control without manual intervention. Architecturally, A$_3$B$_2$ adopts a lightweight asymmetric design inspired by mixture-of-experts with Load Balancing Regularization. Extensive experiments on three few-shot image classification tasks across 11 datasets demonstrate that A$_3$B$_2$ consistently outperforms 11 competitive prompt- and adapter-based baselines.