← Back to all news
arXiv Machine Learning By Bruce Changlong Xu, Jose K. James, Alexander J. Ryu

From Theory to Decision Rule: Calibrating the Noisy-Label Crossover for Vision-Language Model Weak Supervision Across Three Medical-Imaging Benchmarks

Read the original on arXiv Machine Learning →

The study investigates when adding zero‑shot pseudo‑labels from vision‑language models (VLMs) to small expert‑labelled medical imaging datasets ceases to improve, and may even degrade, classifier performance. Across three benchmarks—PatchCamelyon, ISIC 2019, and NIH ChestX‑ray14—the crossover point occurs after roughly 20–50, 100, and 250–500 expert labels, respectively, with additional pseudo‑labels reducing macro‑AUC by up to 0.14 points. The analysis shows that the VLM’s labeling quality, rather than the student model size, limits performance, and a simple bias‑variance rule can predict the benefit of pseudo‑labels in most settings.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

Related stories

arXiv Computer Vision
Aug 27

Label-Free Foundational Model Selection for Medical Image Classification under Distribution Shift via Pseudo Label Discrepancy

The paper introduces a label‑free method called AURCC for selecting the best foundational model for medical image classification when the target domain lacks labels. AURCC uses a pseudo‑label discrepancy computed by the SUDO framework to score models without fine‑tuning. Experiments on chest X‑ray data across three inter‑hospital shifts show that AURCC closely matches the true model ranking, outperforming simple source‑accuracy baselines especially when source data are limited.

By Juan I\~naki Larrea, Lucas Mansilla, Enzo Ferrante
llmscomputer-visionfine-tuningmultimodal
More like this →
arXiv AI
Oct 12

Medical foundation models converge less under label supervision

The study evaluates how well different medical foundation models agree on image embeddings by measuring mutual k‑nearest‑neighbor overlap across 750,000 images from 14 datasets and 18 public models. Results show that label‑supervised models converge the least, sharing only 6–10% of nearest neighbors, whereas models trained with self‑distillation, masked image modeling, or contrastive pretraining share up to 80%. The authors suggest that developers can select pretraining objectives to control convergence and recommend testing unlabeled images for nearest‑neighbor consistency before replacing a model.

By Soroosh Tayebi Arasteh, Sebastian Ziegelmayer, Mahshad Lotfinia, Lisa Adams, Sven Nebelung, Jakob Nikolas Kather, Daniel Truhn
ragefficiency
More like this →