Fine-Grained VLM Fine-tuning via Latent Hierarchical Adapter Learning

TL;DR

This paper introduces LatHAdapter, a hyperbolic space-based fine-tuning method for vision-language models that captures semantic hierarchies to improve few-shot classification, especially for unknown classes.

Contribution

It proposes a novel LatHAdapter that models latent semantic hierarchies using hyperbolic space, addressing limitations of existing adapters in capturing one-to-many category-image associations.

Findings

Outperforms existing fine-tuning methods on four few-shot tasks.

Enhances generalization to unknown classes.

Effectively models semantic hierarchies in hyperbolic space.

Abstract

Adapter-based approaches have garnered attention for fine-tuning pre-trained Vision-Language Models (VLMs) on few-shot classification tasks. These methods strive to develop a lightweight module that better aligns visual and (category) textual representations, thereby enhancing performance on downstream few-shot learning tasks. However, existing adapters generally learn/align (category) textual-visual modalities via explicit spatial proximity in the underlying embedding space, which i) fails to capture the inherent one-to-many associations between categories and image samples and ii) struggles to establish accurate associations between the unknown categories and images. To address these issues, inspired by recent works on hyperbolic learning, we develop a novel Latent Hierarchical Adapter (LatHAdapter) for fine-tuning VLMs on downstream few-shot classification tasks. The core of LatHAdapter is to exploit the latent semantic hierarchy of downstream training data and employ it to provide richer, fine-grained guidance for the adapter learning process. Specifically, LatHAdapter first introduces some learnable `attribute' prompts as the bridge to align categories and images. Then, it projects the categories, attribute prompts, and images within each batch in a hyperbolic space, and employs hierarchical regularization to learn the latent semantic hierarchy of them, thereby fully modeling the inherent one-to-many associations among categories, learnable attributes, and image samples. Extensive experiments on four challenging few-shot tasks show that the proposed LatHAdapter consistently outperforms many other fine-tuning approaches, particularly in adapting known classes and generalizing to unknown classes.