NLM DIR Seminar Schedule
UPCOMING SEMINARS
-
April 22, 2025 Stanley Liang, PhD
Large Vision Model for medical knowledge adaptation -
April 29, 2025 Pascal Mutz
Characterization of covalently closed cirular RNAs detected in (meta)transcriptomic data -
May 2, 2025 Dr. Lang Wu
Integration of multi-omics data in epidemiologic research -
May 6, 2025 Leslie Ronish
TBD -
May 8, 2025 MG Hirsch
TBD
RECENT SEMINARS
-
April 18, 2025 Valentina Boeva, Department of Computer Science, ETH Zurich
Decoding tumor heterogeneity: computational methods for scRNA-seq and spatial omics -
April 8, 2025 Jaya Srivastava
Leveraging a deep learning model to assess the impact of regulatory variants on traits and diseases -
April 1, 2025 Roman Kogay
Horizontal transfer of bacterial operons into eukaryote genomes -
March 25, 2025 Yifan Yang
Adversarial Manipulation and Data Memorization in Large Language Models for Medicine -
March 11, 2025 Sofya Garushyants
Tmn – bacterial anti-phage defense system
Scheduled Seminars on April 22, 2025
In-person: Building 38A/B2N14 NCBI Library or Meeting Link
Contact NLMDIRSeminarScheduling@mail.nih.gov with questions about this seminar.
Abstract:
Large Vision Models (LVMs), built on transformer architectures, have emerged as powerful tools for general-purpose visual representation learning. This study explored the application of LVMs for medical knowledge adaptation, with a focus on automated pneumonia severity prediction from chest X-rays. Leveraging models such as ViT, DINOv2, and BiT, we investigated the roles of LVM as feature encoders under both direct extraction and self-supervised learning settings. Using a dataset of 2,599 chest X-rays labeled with mRALE scores, we evaluate multiple configurations, including Simple Siamese Network based contrastive learning, to assess how LVMs adapt to the medical domain with limited labeled data. The results show that LVMs, particularly when combined with self-supervised learning, can achieve state-of-the-art (SOTA) performance across multiple regression metrics (MSE: 23.83, 95%CI 22.67-25.00; MAE: 3.64, 95%CI 3.54-3.73; explained variance: 0.66, 95%CI 0.63-0.68; R²: 0.65, 95%CI 0.63-0.67; Spearman’s ρ: 0.80, 95%CI 0.77-0.81), when trained on a modest dataset. Notably, while self-supervised learning can enhance general feature quality, it has limited impact on improving alignment with human expert judgment. This work highlights the effectiveness and limitations of LVM-based medical AI pipelines and sets the stage for future efforts involving multi-modality integration and cross-domain adaptation. Ongoing and future work aims to extend this framework through LLM integration and multi-agent learning systems for automated, scalable medical image analysis.