GFL: A Decentralized Federated Learning Framework Based On Blockchain
release_txgrhi65mbeslfn2x42xdxbbeu
by
Yifan Hu, Wei Xia, Jun Xiao, Chao Wu
2020
Abstract
Due to people's emerging concern about data privacy, federated learning(FL)
is currently being widely used. Conventional federated learning uses a highly
centralized architecture, but in a real federated learning scenario, due to the
highly distributed of data nodes and the existence of malicious data nodes, It
is of great challenges for conventional federated learning to improve the
utilization of network bandwidth and maintained the security and robustness of
federated learning under malicious node attacks. In this paper, we propose an
innovative Ring decentralized federated learning algorithm(RDFL) that not only
makes full use of the bandwidth of the network but also improves the security
and robustness of federated learning under malicious node attacks. At the same
time, we encapsulated RDFL into a blochain-based federated learning framework
called Galaxy Federated Learning framework<cit.> and used real data to
perform experiments on the GFL to verify the effectiveness of the GFL.
In text/plain
format
Archived Files and Locations
application/pdf 458.2 kB
file_wwpijeabmvejhln3h4xnaityt4
|
arxiv.org (repository) web.archive.org (webarchive) |
2010.10996v2
access all versions, variants, and formats of this works (eg, pre-prints)