US 11,934,555 B2
Privacy-preserving data curation for federated learning
Youngjin Yoo, Princeton, NJ (US); Gianluca Paladini, Skillman, NJ (US); Eli Gibson, Plainsboro, NJ (US); Pragneshkumar Patel, East Windsor, NJ (US); and Poikavila Ullaskrishnan, Lebanon, NH (US)
Assigned to Siemens Healthineers AG, Forchheim (DE)
Filed by SIEMENS HEALTHINEERS AG, Forchheim (DE)
Filed on Sep. 28, 2021, as Appl. No. 17/449,190.
Prior Publication US 2023/0102732 A1, Mar. 30, 2023
Int. Cl. G06F 21/62 (2013.01); G06N 20/00 (2019.01); G06N 20/20 (2019.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01); G16H 50/70 (2018.01); G16H 80/00 (2018.01)
CPC G06F 21/6245 (2013.01) [G06F 21/6254 (2013.01); G06N 20/20 (2019.01); G16H 50/20 (2018.01); G16H 50/70 (2018.01); G06N 20/00 (2019.01); G16H 30/40 (2018.01); G16H 80/00 (2018.01)] 20 Claims
OG exemplary drawing
 
1. A method for training a model using federated learning by a plurality of collaborator computing devices, the method comprising:
receiving, by a collaborator computing device of the plurality of collaborator computing devices, global model parameters from a parameter aggregation server;
acquiring, by the collaborator computing device, sample data;
transmitting, by the collaborator computing device, an anonymized portion of the sample data to a curation server configured to validate data samples that meet one or more data selection conditions;
receiving, by the collaborator computing device from the curation server, validation for the sample data;
training, by the collaborator computing device, the model using the validated sample data; and
transmitting, by the collaborator computing device, local model parameters for the model to the parameter aggregation server.