NeFut Logo NeFut
Admin Login

[CS.AI] TCLA: Training-Free Class-wise Logit Adaptation for Medical VLMs

Published at: 2026-07-14 22:00 Last updated: 2026-07-15 01:59
#AI #Medical #VLM

Medical Vision-Language Models (VLMs) exhibit strong zero-shot performance, yet their effectiveness still declines on out-of-distribution (OOD) data due to domain shifts and class bias inherited from large-scale pretraining. Existing few-shot adaptation methods typically introduce additional trainable components, which can be unstable in extremely low-data regimes (e.g., 1-shot), and lack robustness on different medical data.

We present TCLA, a purely training-free few-shot adaptation method for Medical VLMs, which is fast and model-agnostic. TCLA corrects inference logits based on a small set of support samples, boosting pretrained VLMs performance by improving inter-class deconfusion and reducing domain shift.

Extensive experiments on nine datasets across multiple medical imaging modalities including X-ray, Ultrasound, MRI, CT, Histopathology, demonstrate that TCLA consistently improves OOD performance of Medical VLMs and, in most cases, outperforms existing training-based adaptation methods.

Blogger's Review: The introduction of the TCLA method provides an efficient adaptation strategy for vision-language models in the medical domain, particularly under data scarcity. This approach not only enhances model accuracy but also offers new insights for future AI research in medicine. Its model-agnostic nature allows for broad applicability across various scenarios.

Original Source: https://arxiv.org/abs/2607.09562

[h] Back to Home