Multi Site Data

Multi-site data analysis focuses on leveraging data from multiple sources to improve the generalizability, robustness, and statistical power of scientific studies, while addressing privacy concerns. Current research emphasizes federated learning approaches, employing techniques like multiply-robust estimation, noise-resilient training, and Bayesian networks, to combine data efficiently and account for site-specific variations in data distribution and annotation quality. This methodology is proving valuable across diverse fields, including medical imaging (e.g., brain MRI segmentation and lesion detection) and solar forecasting, enabling more accurate and reliable models that can be applied across different institutions and populations.

Papers

September 22, 2023

Multiply Robust Federated Estimation of Targeted Average Treatment Effects
Larry Han, Zhu Shen, Jose Zubizarreta
Covariate Shift Multi Center Study Average Treatment Effect Robust Federated Learning Multi Site Data

August 31, 2023

Improving Multiple Sclerosis Lesion Segmentation Across Clinical Sites: A Federated Learning Approach with Noise-Resilient Training
Lei Bai, Dongang Wang, Michael Barnett, Mariano Cabezas, Weidong Cai, Fernando Calamante, Kain Kyle, Dongnan Liu, Linda Ly, Aria Nguyen, Chun-Chien Shieh, Ryan Sullivan, Hengrui Wang, Geng Zhan, Wanli Ouyang, Chenyu Wang
Federated Learning Multiple Sclerosis Federated Learning Approach Label Correction Multi Site Data Clinical Trial Site

January 6, 2023

Learning Personalized Brain Functional Connectivity of MDD Patients from Multiple Sites via Federated Bayesian Networks
Shuai Liu, Xiao Guo, Shun Qi, Huaning Wang, Xiangyu Chang
Functional Magnetic Resonance Imaging Joint Estimation Multi Site Major Depressive Disorder Multi Site Data Functional Network Connectivity

November 4, 2022

Fighting the scanner effect in brain MRI segmentation with a progressive level-of-detail network trained on multi-site data
Michele Svanera, Mattia Savardi, Alberto Signoroni, Sergio Benini, Lars Muckli
Domain Shift 3D Convolutional Brain Segmentation Intensity Distribution Brain Data Multi Site Data Level of Detail

November 3, 2022

Sky-image-based solar forecasting using deep learning with multi-location data: training models locally, globally or via transfer learning?
Yuhao Nie, Quentin Paletta, Andea Scott, Luis Martin Pomares, Guillaume Arbod, Sgouris Sgouridis, Joan Lasenby, Adam Brandt
Deep Learning Transfer Learning Training Model Solar Forecasting Sky Image Datasets Multi Site Data

May 31, 2022

FedHarmony: Unlearning Scanner Bias with Distributed Data
Nicola K Dinsdale, Mark Jenkinson, Ana IL Namburete
Data Distribution Image Harmonization Data Harmonization Multi Site Data

November 7, 2021

Acquisition-invariant brain MRI segmentation with informative uncertainties
Pedro Borges, Richard Shaw, Thomas Varsavsky, Kerstin Klaser, David Thomas, Ivana Drobnjak, Sebastien Ourselin, M Jorge Cardoso
Relevant Covariates Brain Segmentation Uncertainty Information Sequence Design Butterfly Effect Multi Site Data

Multi Site Data

Papers

Multiply Robust Federated Estimation of Targeted Average Treatment Effects

Improving Multiple Sclerosis Lesion Segmentation Across Clinical Sites: A Federated Learning Approach with Noise-Resilient Training

Learning Personalized Brain Functional Connectivity of MDD Patients from Multiple Sites via Federated Bayesian Networks

Fighting the scanner effect in brain MRI segmentation with a progressive level-of-detail network trained on multi-site data

Sky-image-based solar forecasting using deep learning with multi-location data: training models locally, globally or via transfer learning?

FedHarmony: Unlearning Scanner Bias with Distributed Data

Acquisition-invariant brain MRI segmentation with informative uncertainties