CoDoL: Conditional Domain Prompt Learning for Out-of-Distribution Generalization

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zhang, Min, Jiang, Bo, Zhou, Jie, Liu, Yimeng, Lin, Xin
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908546440888320
author Zhang, Min
Jiang, Bo
Zhou, Jie
Liu, Yimeng
Lin, Xin
author_facet Zhang, Min
Jiang, Bo
Zhou, Jie
Liu, Yimeng
Lin, Xin
contents Recent advances in pre-training vision-language models (VLMs), e.g., contrastive language-image pre-training (CLIP) methods, have shown great potential in learning out-of-distribution (OOD) representations. Despite showing competitive performance, the prompt-based CLIP methods still suffer from: i) inaccurate text descriptions, which leads to degraded accuracy and robustness, and poses a challenge for zero-shot CLIP methods. ii) limited vision-language embedding alignment, which significantly affects the generalization performance. To tackle the above issues, this paper proposes a novel Conditional Domain prompt Learning (CoDoL) method, which utilizes readily-available domain information to form prompts and improves the vision-language embedding alignment for improving OOD generalization. To capture both instance-specific and domain-specific information, we further propose a lightweight Domain Meta Network (DMN) to generate input-conditional tokens for images in each domain. Extensive experiments on four OOD benchmarks (PACS, VLCS, OfficeHome and DigitDG) validate the effectiveness of our proposed CoDoL in terms of improving the vision-language embedding alignment as well as the out-of-distribution generalization performance.
format Preprint
id arxiv_https___arxiv_org_abs_2509_15330
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CoDoL: Conditional Domain Prompt Learning for Out-of-Distribution Generalization
Zhang, Min
Jiang, Bo
Zhou, Jie
Liu, Yimeng
Lin, Xin
Computer Vision and Pattern Recognition
Recent advances in pre-training vision-language models (VLMs), e.g., contrastive language-image pre-training (CLIP) methods, have shown great potential in learning out-of-distribution (OOD) representations. Despite showing competitive performance, the prompt-based CLIP methods still suffer from: i) inaccurate text descriptions, which leads to degraded accuracy and robustness, and poses a challenge for zero-shot CLIP methods. ii) limited vision-language embedding alignment, which significantly affects the generalization performance. To tackle the above issues, this paper proposes a novel Conditional Domain prompt Learning (CoDoL) method, which utilizes readily-available domain information to form prompts and improves the vision-language embedding alignment for improving OOD generalization. To capture both instance-specific and domain-specific information, we further propose a lightweight Domain Meta Network (DMN) to generate input-conditional tokens for images in each domain. Extensive experiments on four OOD benchmarks (PACS, VLCS, OfficeHome and DigitDG) validate the effectiveness of our proposed CoDoL in terms of improving the vision-language embedding alignment as well as the out-of-distribution generalization performance.
title CoDoL: Conditional Domain Prompt Learning for Out-of-Distribution Generalization
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.15330