ProAPO: Progressively Automatic Prompt Optimization for Visual Classification

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Qu, Xiangyan, Gou, Gaopeng, Zhuang, Jiamin, Yu, Jing, Song, Kun, Wang, Qihao, Li, Yili, Xiong, Gang
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911438804615168
author Qu, Xiangyan
Gou, Gaopeng
Zhuang, Jiamin
Yu, Jing
Song, Kun
Wang, Qihao
Li, Yili
Xiong, Gang
author_facet Qu, Xiangyan
Gou, Gaopeng
Zhuang, Jiamin
Yu, Jing
Song, Kun
Wang, Qihao
Li, Yili
Xiong, Gang
contents Vision-language models (VLMs) have made significant progress in image classification by training with large-scale paired image-text data. Their performances largely depend on the prompt quality. While recent methods show that visual descriptions generated by large language models (LLMs) enhance the generalization of VLMs, class-specific prompts may be inaccurate or lack discrimination due to the hallucination in LLMs. In this paper, we aim to find visually discriminative prompts for fine-grained categories with minimal supervision and no human-in-the-loop. An evolution-based algorithm is proposed to progressively optimize language prompts from task-specific templates to class-specific descriptions. Unlike optimizing templates, the search space shows an explosion in class-specific candidate prompts. This increases prompt generation costs, iterative times, and the overfitting problem. To this end, we first introduce several simple yet effective edit-based and evolution-based operations to generate diverse candidate prompts by one-time query of LLMs. Then, two sampling strategies are proposed to find a better initial search point and reduce traversed categories, saving iteration costs. Moreover, we apply a novel fitness score with entropy constraints to mitigate overfitting. In a challenging one-shot image classification setting, our method outperforms existing textual prompt-based methods and improves LLM-generated description methods across 13 datasets. Meanwhile, we demonstrate that our optimal prompts improve adapter-based methods and transfer effectively across different backbones.
format Preprint
id arxiv_https___arxiv_org_abs_2502_19844
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle ProAPO: Progressively Automatic Prompt Optimization for Visual Classification
Qu, Xiangyan
Gou, Gaopeng
Zhuang, Jiamin
Yu, Jing
Song, Kun
Wang, Qihao
Li, Yili
Xiong, Gang
Computer Vision and Pattern Recognition
Vision-language models (VLMs) have made significant progress in image classification by training with large-scale paired image-text data. Their performances largely depend on the prompt quality. While recent methods show that visual descriptions generated by large language models (LLMs) enhance the generalization of VLMs, class-specific prompts may be inaccurate or lack discrimination due to the hallucination in LLMs. In this paper, we aim to find visually discriminative prompts for fine-grained categories with minimal supervision and no human-in-the-loop. An evolution-based algorithm is proposed to progressively optimize language prompts from task-specific templates to class-specific descriptions. Unlike optimizing templates, the search space shows an explosion in class-specific candidate prompts. This increases prompt generation costs, iterative times, and the overfitting problem. To this end, we first introduce several simple yet effective edit-based and evolution-based operations to generate diverse candidate prompts by one-time query of LLMs. Then, two sampling strategies are proposed to find a better initial search point and reduce traversed categories, saving iteration costs. Moreover, we apply a novel fitness score with entropy constraints to mitigate overfitting. In a challenging one-shot image classification setting, our method outperforms existing textual prompt-based methods and improves LLM-generated description methods across 13 datasets. Meanwhile, we demonstrate that our optimal prompts improve adapter-based methods and transfer effectively across different backbones.
title ProAPO: Progressively Automatic Prompt Optimization for Visual Classification
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.19844