Robust Test-time Video-Text Retrieval: Benchmarking and Adapting for Query Shifts

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Zhang, Bingqing, Cao, Zhuo, Du, Heming, Li, Yang, Li, Xue, Liu, Jiajun, Wang, Sen
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918462822023168
author Zhang, Bingqing
Cao, Zhuo
Du, Heming
Li, Yang
Li, Xue
Liu, Jiajun
Wang, Sen
author_facet Zhang, Bingqing
Cao, Zhuo
Du, Heming
Li, Yang
Li, Xue
Liu, Jiajun
Wang, Sen
contents Modern video-text retrieval (VTR) models excel on in-distribution benchmarks but are highly vulnerable to real-world query shifts, where the distribution of query data deviates from the training domain, leading to a sharp performance drop. Existing image-focused robustness solutions are inadequate to handle this vulnerability in video, as they fail to address the complex spatio-temporal dynamics inherent in these shifts. To systematically evaluate this vulnerability, we first introduce a comprehensive benchmark featuring 12 distinct types of video perturbations across five severity degrees. Analysis on this benchmark reveals that query shifts amplify the hubness phenomenon, where a few gallery items become dominant "hubs" that attract a disproportionate number of queries. To mitigate this, we then propose HAT-VTR (Hubness Alleviation for Test-time Video-Text Retrieval), as our baseline test-time adaptation framework designed to directly counteract hubness in VTR. It leverages two key components: a Hubness Suppression Memory to refine similarity scores, and multi-granular losses to enforce temporal feature consistency. Extensive experiments demonstrate that HAT-VTR substantially improves robustness, consistently outperforming prior methods across diverse query shift scenarios, and enhancing model reliability for real-world applications.
format Preprint
id arxiv_https___arxiv_org_abs_2604_20851
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Robust Test-time Video-Text Retrieval: Benchmarking and Adapting for Query Shifts
Zhang, Bingqing
Cao, Zhuo
Du, Heming
Li, Yang
Li, Xue
Liu, Jiajun
Wang, Sen
Information Retrieval
Artificial Intelligence
Computer Vision and Pattern Recognition
68T05
I.2.10; H.3.3
Modern video-text retrieval (VTR) models excel on in-distribution benchmarks but are highly vulnerable to real-world query shifts, where the distribution of query data deviates from the training domain, leading to a sharp performance drop. Existing image-focused robustness solutions are inadequate to handle this vulnerability in video, as they fail to address the complex spatio-temporal dynamics inherent in these shifts. To systematically evaluate this vulnerability, we first introduce a comprehensive benchmark featuring 12 distinct types of video perturbations across five severity degrees. Analysis on this benchmark reveals that query shifts amplify the hubness phenomenon, where a few gallery items become dominant "hubs" that attract a disproportionate number of queries. To mitigate this, we then propose HAT-VTR (Hubness Alleviation for Test-time Video-Text Retrieval), as our baseline test-time adaptation framework designed to directly counteract hubness in VTR. It leverages two key components: a Hubness Suppression Memory to refine similarity scores, and multi-granular losses to enforce temporal feature consistency. Extensive experiments demonstrate that HAT-VTR substantially improves robustness, consistently outperforming prior methods across diverse query shift scenarios, and enhancing model reliability for real-world applications.
title Robust Test-time Video-Text Retrieval: Benchmarking and Adapting for Query Shifts
topic Information Retrieval
Artificial Intelligence
Computer Vision and Pattern Recognition
68T05
I.2.10; H.3.3
url https://arxiv.org/abs/2604.20851