Data-Free Distillation Improves Efficiency and Privacy in Federated Thorax Disease Analysis. Li, M. & Yang, G. In IEEE EMBS International Conference on Data Science and Engineering in Healthcare, Medicine & Biology, 2023. ![Data-Free Distillation Improves Efficiency and Privacy in Federated Thorax Disease Analysis [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper ![Data-Free Distillation Improves Efficiency and Privacy in Federated Thorax Disease Analysis [link]](https://bibbase.org/img/filetypes/link.svg "link")
Website abstract bibtex Thorax disease analysis in large-scale, multi-centre, and multi-scanner settings is often limited by strict privacy policies. Federated learning (FL) offers a potential solution, while traditional parameter-based FL can be limited by issues such as high communication costs, data leakage, and heterogeneity. Distillation-based FL can improve efficiency, but it relies on a proxy dataset, which is often impractical in clinical practice. To address these challenges, we introduce a data-free distillation-based FL approach FedKDF. In FedKDF, the server employs a lightweight generator to aggregate knowledge from different clients without requiring access to their private data or a proxy dataset. FedKDF combines the predictors from clients into a single, unified predictor, which is further optimized using the learned knowledge in the lightweight generator. Our empirical experiments demonstrate that FedKDF offers a robust solution for efficient, privacy-preserving federated thorax disease analysis.
@inproceedings{
title = {Data-Free Distillation Improves Efficiency and Privacy in Federated Thorax Disease Analysis},
type = {inproceedings},
year = {2023},
issue = {3},
websites = {http://arxiv.org/abs/2310.18346},
id = {f570dceb-25f3-38d1-b5ba-99dba69fb031},
created = {2024-01-13T07:02:57.026Z},
file_attached = {true},
profile_id = {d5b9f595-4104-33d2-b640-e84dee1b7ded},
last_modified = {2024-01-13T08:14:19.397Z},
read = {false},
starred = {false},
authored = {true},
confirmed = {true},
hidden = {false},
private_publication = {false},
abstract = {Thorax disease analysis in large-scale, multi-centre, and multi-scanner settings is often limited by strict privacy policies. Federated learning (FL) offers a potential solution, while traditional parameter-based FL can be limited by issues such as high communication costs, data leakage, and heterogeneity. Distillation-based FL can improve efficiency, but it relies on a proxy dataset, which is often impractical in clinical practice. To address these challenges, we introduce a data-free distillation-based FL approach FedKDF. In FedKDF, the server employs a lightweight generator to aggregate knowledge from different clients without requiring access to their private data or a proxy dataset. FedKDF combines the predictors from clients into a single, unified predictor, which is further optimized using the learned knowledge in the lightweight generator. Our empirical experiments demonstrate that FedKDF offers a robust solution for efficient, privacy-preserving federated thorax disease analysis.},
bibtype = {inproceedings},
author = {Li, Ming and Yang, Guang},
booktitle = {IEEE EMBS International Conference on Data Science and Engineering in Healthcare, Medicine & Biology}
}
Downloads: 0
{"_id":"sJCo64Mvs7TKd7T2m","bibbaseid":"li-yang-datafreedistillationimprovesefficiencyandprivacyinfederatedthoraxdiseaseanalysis-2023","author_short":["Li, M.","Yang, G."],"bibdata":{"title":"Data-Free Distillation Improves Efficiency and Privacy in Federated Thorax Disease Analysis","type":"inproceedings","year":"2023","issue":"3","websites":"http://arxiv.org/abs/2310.18346","id":"f570dceb-25f3-38d1-b5ba-99dba69fb031","created":"2024-01-13T07:02:57.026Z","file_attached":"true","profile_id":"d5b9f595-4104-33d2-b640-e84dee1b7ded","last_modified":"2024-01-13T08:14:19.397Z","read":false,"starred":false,"authored":"true","confirmed":"true","hidden":false,"private_publication":false,"abstract":"Thorax disease analysis in large-scale, multi-centre, and multi-scanner settings is often limited by strict privacy policies. Federated learning (FL) offers a potential solution, while traditional parameter-based FL can be limited by issues such as high communication costs, data leakage, and heterogeneity. Distillation-based FL can improve efficiency, but it relies on a proxy dataset, which is often impractical in clinical practice. To address these challenges, we introduce a data-free distillation-based FL approach FedKDF. In FedKDF, the server employs a lightweight generator to aggregate knowledge from different clients without requiring access to their private data or a proxy dataset. FedKDF combines the predictors from clients into a single, unified predictor, which is further optimized using the learned knowledge in the lightweight generator. Our empirical experiments demonstrate that FedKDF offers a robust solution for efficient, privacy-preserving federated thorax disease analysis.","bibtype":"inproceedings","author":"Li, Ming and Yang, Guang","booktitle":"IEEE EMBS International Conference on Data Science and Engineering in Healthcare, Medicine & Biology","bibtex":"@inproceedings{\n title = {Data-Free Distillation Improves Efficiency and Privacy in Federated Thorax Disease Analysis},\n type = {inproceedings},\n year = {2023},\n issue = {3},\n websites = {http://arxiv.org/abs/2310.18346},\n id = {f570dceb-25f3-38d1-b5ba-99dba69fb031},\n created = {2024-01-13T07:02:57.026Z},\n file_attached = {true},\n profile_id = {d5b9f595-4104-33d2-b640-e84dee1b7ded},\n last_modified = {2024-01-13T08:14:19.397Z},\n read = {false},\n starred = {false},\n authored = {true},\n confirmed = {true},\n hidden = {false},\n private_publication = {false},\n abstract = {Thorax disease analysis in large-scale, multi-centre, and multi-scanner settings is often limited by strict privacy policies. Federated learning (FL) offers a potential solution, while traditional parameter-based FL can be limited by issues such as high communication costs, data leakage, and heterogeneity. Distillation-based FL can improve efficiency, but it relies on a proxy dataset, which is often impractical in clinical practice. To address these challenges, we introduce a data-free distillation-based FL approach FedKDF. In FedKDF, the server employs a lightweight generator to aggregate knowledge from different clients without requiring access to their private data or a proxy dataset. FedKDF combines the predictors from clients into a single, unified predictor, which is further optimized using the learned knowledge in the lightweight generator. Our empirical experiments demonstrate that FedKDF offers a robust solution for efficient, privacy-preserving federated thorax disease analysis.},\n bibtype = {inproceedings},\n author = {Li, Ming and Yang, Guang},\n booktitle = {IEEE EMBS International Conference on Data Science and Engineering in Healthcare, Medicine & Biology}\n}","author_short":["Li, M.","Yang, G."],"urls":{"Paper":"https://bibbase.org/service/mendeley/d5b9f595-4104-33d2-b640-e84dee1b7ded/file/56054835-e67f-6699-5c36-afab400dbdcf/231018346.pdf.pdf","Website":"http://arxiv.org/abs/2310.18346"},"biburl":"https://bibbase.org/service/mendeley/d5b9f595-4104-33d2-b640-e84dee1b7ded","bibbaseid":"li-yang-datafreedistillationimprovesefficiencyandprivacyinfederatedthoraxdiseaseanalysis-2023","role":"author","metadata":{"authorlinks":{}},"downloads":0},"bibtype":"inproceedings","biburl":"https://bibbase.org/service/mendeley/d5b9f595-4104-33d2-b640-e84dee1b7ded","dataSources":["2252seNhipfTmjEBQ"],"keywords":[],"search_terms":["data","free","distillation","improves","efficiency","privacy","federated","thorax","disease","analysis","li","yang"],"title":"Data-Free Distillation Improves Efficiency and Privacy in Federated Thorax Disease Analysis","year":2023}