RingFFL: A Ring-Architecture-Based Fair Federated Learning Framework
In the ring-architecture-based federated learning framework, security and fairness are severely compromised when dishonest clients abort the training process after obtaining useful information. To solve the problem, we propose a <b>Ring-</b> architecture-based <b>F</b>air <...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-02-01
|
| Series: | Future Internet |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1999-5903/15/2/68 |
| _version_ | 1797620921141297152 |
|---|---|
| author | Lu Han Xiaohong Huang Dandan Li Yong Zhang |
| author_facet | Lu Han Xiaohong Huang Dandan Li Yong Zhang |
| author_sort | Lu Han |
| collection | DOAJ |
| description | In the ring-architecture-based federated learning framework, security and fairness are severely compromised when dishonest clients abort the training process after obtaining useful information. To solve the problem, we propose a <b>Ring-</b> architecture-based <b>F</b>air <b>F</b>ederated <b>L</b>earning framework called RingFFL, in which we design a penalty mechanism for FL. Before the training starts in each round, all clients that will participate in the training pay deposits in a set order and record the transactions on the blockchain to ensure that they are not tampered with. Subsequently, the clients perform the FL training process, and the correctness of the models transmitted by the clients is guaranteed by the HASH algorithm during the training process. When all clients perform honestly, each client can obtain the final model, and the number of digital currencies in each client’s wallet is kept constant; otherwise, the deposits of clients who leave halfway will be compensated to the clients who perform honestly during the training process. In this way, through the penalty mechanism, all clients either obtain the final model or are compensated, thus ensuring the fairness of federated learning. The security analysis and experimental results show that RingFFL not only guarantees the accuracy and security of the federated learning model but also guarantees the fairness. |
| first_indexed | 2024-03-11T08:48:25Z |
| format | Article |
| id | doaj.art-f4bc5cee9a864f25a7f53ee6b3b8c217 |
| institution | Directory Open Access Journal |
| issn | 1999-5903 |
| language | English |
| last_indexed | 2024-03-11T08:48:25Z |
| publishDate | 2023-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Future Internet |
| spelling | doaj.art-f4bc5cee9a864f25a7f53ee6b3b8c2172023-11-16T20:37:50ZengMDPI AGFuture Internet1999-59032023-02-011526810.3390/fi15020068RingFFL: A Ring-Architecture-Based Fair Federated Learning FrameworkLu Han0Xiaohong Huang1Dandan Li2Yong Zhang3School of Computer Science (National Pilot Software Engineering School), University of Posts and Telecommunication, Beijing 100876, ChinaSchool of Computer Science (National Pilot Software Engineering School), University of Posts and Telecommunication, Beijing 100876, ChinaSchool of Computer Science (National Pilot Software Engineering School), University of Posts and Telecommunication, Beijing 100876, ChinaZhongguancun Laboratory, Beijing 100094, ChinaIn the ring-architecture-based federated learning framework, security and fairness are severely compromised when dishonest clients abort the training process after obtaining useful information. To solve the problem, we propose a <b>Ring-</b> architecture-based <b>F</b>air <b>F</b>ederated <b>L</b>earning framework called RingFFL, in which we design a penalty mechanism for FL. Before the training starts in each round, all clients that will participate in the training pay deposits in a set order and record the transactions on the blockchain to ensure that they are not tampered with. Subsequently, the clients perform the FL training process, and the correctness of the models transmitted by the clients is guaranteed by the HASH algorithm during the training process. When all clients perform honestly, each client can obtain the final model, and the number of digital currencies in each client’s wallet is kept constant; otherwise, the deposits of clients who leave halfway will be compensated to the clients who perform honestly during the training process. In this way, through the penalty mechanism, all clients either obtain the final model or are compensated, thus ensuring the fairness of federated learning. The security analysis and experimental results show that RingFFL not only guarantees the accuracy and security of the federated learning model but also guarantees the fairness.https://www.mdpi.com/1999-5903/15/2/68federated learningfairnessblockchainring architecture |
| spellingShingle | Lu Han Xiaohong Huang Dandan Li Yong Zhang RingFFL: A Ring-Architecture-Based Fair Federated Learning Framework Future Internet federated learning fairness blockchain ring architecture |
| title | RingFFL: A Ring-Architecture-Based Fair Federated Learning Framework |
| title_full | RingFFL: A Ring-Architecture-Based Fair Federated Learning Framework |
| title_fullStr | RingFFL: A Ring-Architecture-Based Fair Federated Learning Framework |
| title_full_unstemmed | RingFFL: A Ring-Architecture-Based Fair Federated Learning Framework |
| title_short | RingFFL: A Ring-Architecture-Based Fair Federated Learning Framework |
| title_sort | ringffl a ring architecture based fair federated learning framework |
| topic | federated learning fairness blockchain ring architecture |
| url | https://www.mdpi.com/1999-5903/15/2/68 |
| work_keys_str_mv | AT luhan ringfflaringarchitecturebasedfairfederatedlearningframework AT xiaohonghuang ringfflaringarchitecturebasedfairfederatedlearningframework AT dandanli ringfflaringarchitecturebasedfairfederatedlearningframework AT yongzhang ringfflaringarchitecturebasedfairfederatedlearningframework |