BioVLM: Routing Prompts, Not Parameters, for Cross-Modality Generalization in Biomedical VLMs
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_ | 1866917420737757184 |
|---|---|
| author | Singha, Mainak Gupta, Tanisha Jha, Ankit Khan, Muhammad Haris Ghosh, Sayantani Banerjee, Biplab |
| author_facet | Singha, Mainak Gupta, Tanisha Jha, Ankit Khan, Muhammad Haris Ghosh, Sayantani Banerjee, Biplab |
| contents | Pretrained biomedical vision-language models (VLMs) such as BioMedCLIP perform well on average but often degrade on challenging modalities where inter-class margins are small and acquisition-specific variations are pronounced, especially under few-shot supervision and when modality priors differ from pretraining corpora substantially. We propose BioVLM, a prompt-learning framework that improves cross-domain generalization without extensive backbone fine-tuning. BioVLM learns a diverse prompt bank and introduces dynamic prompt selection: for each input, it selects the most discriminative prompts via a low-entropy criterion on the predictive distribution, effectively coupling sparse few-shot evidence with rich LLM semantic priors. To strengthen this coupling, we distill high-confidence LLM-derived attributes and enforce robust knowledge transfer through strong/weak augmentation consistency. At test time, BioVLM adapts by choosing modality-appropriate prompts, enabling transfer to unseen categories and domains, while keeping training lightweight and inference efficient. On 11 MedMNIST+ 2D datasets, BioVLM achieves new state of the art across three distinct generalization settings. Codes are available at https://github.com/mainaksingha01/BioVLM. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_17629 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | BioVLM: Routing Prompts, Not Parameters, for Cross-Modality Generalization in Biomedical VLMs Singha, Mainak Gupta, Tanisha Jha, Ankit Khan, Muhammad Haris Ghosh, Sayantani Banerjee, Biplab Computer Vision and Pattern Recognition Pretrained biomedical vision-language models (VLMs) such as BioMedCLIP perform well on average but often degrade on challenging modalities where inter-class margins are small and acquisition-specific variations are pronounced, especially under few-shot supervision and when modality priors differ from pretraining corpora substantially. We propose BioVLM, a prompt-learning framework that improves cross-domain generalization without extensive backbone fine-tuning. BioVLM learns a diverse prompt bank and introduces dynamic prompt selection: for each input, it selects the most discriminative prompts via a low-entropy criterion on the predictive distribution, effectively coupling sparse few-shot evidence with rich LLM semantic priors. To strengthen this coupling, we distill high-confidence LLM-derived attributes and enforce robust knowledge transfer through strong/weak augmentation consistency. At test time, BioVLM adapts by choosing modality-appropriate prompts, enabling transfer to unseen categories and domains, while keeping training lightweight and inference efficient. On 11 MedMNIST+ 2D datasets, BioVLM achieves new state of the art across three distinct generalization settings. Codes are available at https://github.com/mainaksingha01/BioVLM. |
| title | BioVLM: Routing Prompts, Not Parameters, for Cross-Modality Generalization in Biomedical VLMs |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2604.17629 |