Leveraging Vision-Language Models as Weak Annotators in Active Learning
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917452891291648 |
|---|---|
| author | Nguyen, Phuong Ngoc Shiku, Kaito Bise, Ryoma Uchida, Seiichi Matsuo, Shinnosuke |
| author_facet | Nguyen, Phuong Ngoc Shiku, Kaito Bise, Ryoma Uchida, Seiichi Matsuo, Shinnosuke |
| contents | Active learning aims to reduce annotation cost by selectively querying informative samples for supervision under a limited labeling budget. In this work, we investigate how vision-language models (VLMs) can be leveraged to further reduce the reliance on costly human annotation within the active learning paradigm. To this end, we find that the reliability of VLMs varies significantly with label granularity in fine-grained recognition tasks: they perform poorly on fine-grained labels but can provide accurate coarse-grained labels. Leveraging this property, we propose an active learning framework that combines fine-grained human annotations with coarse-grained VLM-generated weak labels through instance-wise label assignment. We further model the systematic noise in VLM-generated labels using a small set of trusted full labels. Experiments on CUB200 and FGVC-Aircraft show that the proposed framework consistently outperforms existing active learning methods under the same annotation budget. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_00480 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Leveraging Vision-Language Models as Weak Annotators in Active Learning Nguyen, Phuong Ngoc Shiku, Kaito Bise, Ryoma Uchida, Seiichi Matsuo, Shinnosuke Computer Vision and Pattern Recognition Active learning aims to reduce annotation cost by selectively querying informative samples for supervision under a limited labeling budget. In this work, we investigate how vision-language models (VLMs) can be leveraged to further reduce the reliance on costly human annotation within the active learning paradigm. To this end, we find that the reliability of VLMs varies significantly with label granularity in fine-grained recognition tasks: they perform poorly on fine-grained labels but can provide accurate coarse-grained labels. Leveraging this property, we propose an active learning framework that combines fine-grained human annotations with coarse-grained VLM-generated weak labels through instance-wise label assignment. We further model the systematic noise in VLM-generated labels using a small set of trusted full labels. Experiments on CUB200 and FGVC-Aircraft show that the proposed framework consistently outperforms existing active learning methods under the same annotation budget. |
| title | Leveraging Vision-Language Models as Weak Annotators in Active Learning |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.00480 |