AI Models Close to your Chest: Robust Federated Learning Strategies for Multi-site CT
Authors:
Edward H. Lee,
Brendan Kelly,
Emre Altinmakas,
Hakan Dogan,
Maryam Mohammadzadeh,
Errol Colak,
Steve Fu,
Olivia Choudhury,
Ujjwal Ratan,
Felipe Kitamura,
Hernan Chaves,
Jimmy Zheng,
Mourad Said,
Eduardo Reis,
Jaekwang Lim,
Patricia Yokoo,
Courtney Mitchell,
Golnaz Houshmand,
Marzyeh Ghassemi,
Ronan Killeen,
Wendy Qiu,
Joel Hayden,
Farnaz Rafiee,
Chad Klochko,
Nicholas Bevins
, et al. (5 additional authors not shown)
Abstract:
While it is well known that population differences from genetics, sex, race, and environmental factors contribute to disease, AI studies in medicine have largely focused on locoregional patient cohorts with less diverse data sources. Such limitation stems from barriers to large-scale data share and ethical concerns over data privacy. Federated learning (FL) is one potential pathway for AI developm…
▽ More
While it is well known that population differences from genetics, sex, race, and environmental factors contribute to disease, AI studies in medicine have largely focused on locoregional patient cohorts with less diverse data sources. Such limitation stems from barriers to large-scale data share and ethical concerns over data privacy. Federated learning (FL) is one potential pathway for AI development that enables learning across hospitals without data share. In this study, we show the results of various FL strategies on one of the largest and most diverse COVID-19 chest CT datasets: 21 participating hospitals across five continents that comprise >10,000 patients with >1 million images. We also propose an FL strategy that leverages synthetically generated data to overcome class and size imbalances. We also describe the sources of data heterogeneity in the context of FL, and show how even among the correctly labeled populations, disparities can arise due to these biases.
△ Less
Submitted 13 April, 2023; v1 submitted 23 March, 2023;
originally announced March 2023.
Use of Power Flow Controllers to Enhance Transmission Network Utilisation on the Irish Transmission Network
Authors:
A. Soroudi,
B. Oconnell,
J. Kelliher,
J. Osullivan,
B. Kelly,
M. Walsh,
F. Kreikebaum,
Y. Meyer
Abstract:
The Ireland and Northern Ireland power system is in a period of rapid transition from conventional generation to renewable generation and has seen a rapid increase in large demand sites requiring connection into the backbone transmission system. The role of EirGrid and SONI as Transmission System Operator in Ireland and Northern Ireland is to operate, maintain and develop the electricity transmiss…
▽ More
The Ireland and Northern Ireland power system is in a period of rapid transition from conventional generation to renewable generation and has seen a rapid increase in large demand sites requiring connection into the backbone transmission system. The role of EirGrid and SONI as Transmission System Operator in Ireland and Northern Ireland is to operate, maintain and develop the electricity transmission network. EirGrid ensures that new transmission projects are developed in a way that balances technical, economic, community and other stakeholder considerations. This has resulted in much more detailed evaluation of planning options to maximize utilisation of the existing network which may include but are not limited to uprating the capacity of the existing transmission system mainly through the thermal or voltage uprate of existing circuits. Another method to increase network utilisation is to strategically deploy Power Flow Control devices to relieve system overloads and maximise network transfer capacities. EirGrid has developed a new grid development strategy that places particular emphasis on identifying technologies to help resolve network issues. This paper presents the study findings for the application of a power flow controller (PFC) to relieve system issues.
△ Less
Submitted 15 November, 2019;
originally announced November 2019.