Article

Revealing the Impact of Pre-training Data on Medical Foundation Models

2025-04-03

Abstract excerpt

<title>Abstract</title> <p>Medical foundation models (FM), pre-trained on large-scale unlabelled data, have demonstrated robust performance and high efficiency when fine-tuned to various clinically relevant applications. However, the impact of pre-training data on medical FM performance such as generalisability and fairness, which form the foundation in fine-tuned models, remains unexplored. To address this, we s...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 0ab5ff84-5add-56e8-bb22-93fee6cdb915
DOI: 10.21203/rs.3.rs-6080254/v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Revealing the Impact of Pre-training Data on Medical Foundation ModelsDOI 10.21203/rs.3.rs-6080254/v1Read

Comparison of foundation models and transfer learning strategies for diabetic retinopathy classificationSemantic proximity Read

UWF-FM: A Foundation Model for Comprehensive Ultra-Widefield Retinal AssessmentSemantic proximity Read

An Efficient and Interpretable Foundation Model for Retinal Image Analysis in Disease DiagnosisSemantic proximity Read

Enhancing Glaucoma Detection through Supervised Pre-training with Intermediate Phenotypes: A Multi-Institutional StudySemantic proximity Read

CausalFund: Causality-Inspired Domain Generalization in Retinal Fundus Imaging for Low-Resource ScreeningSemantic proximity Read

A Federated Learning-based Optic Disc and Cup Segmentation Model for Glaucoma Monitoring In Color Fundus PhotographsSemantic proximity Read

Expertise-informed Generative AI Enables Ultra-High Data Efficiency for Building Generalist Medical Foundation ModelSemantic proximity Read

Contrastive learning-based pretraining improves representation and transferability of diabetic retinopathy classification modelsSemantic proximity Read

In-context learning for data-efficient classification of diabetic retinopathy with multimodal foundation modelsSemantic proximity Read

Comparison of Foundation and Supervised Learning-Based Models for Detection of Referable Glaucoma from Fundus PhotographsSemantic proximity Read

Self-supervised retinal thickness prediction enables deep learning from unlabeled data to boost classification of diabetic retinopathySemantic proximity Read

Comparing tangible retinal image characteristics with deep learning features reveals their complementarity for gene association and disease predictionSemantic proximity Read