Blockchain-based Node-aware Dynamic Weighting Methods for Improving Federated Learning Performance
-
- You Jun Kim
- Kyung Hee University
-
- Choong Seon Hong
- Kyung Hee University
抄録
Federated learning (FL) is a decentralized learning method that deviated from the conventional centralized learning. The FL progresses learning locally on each device and gradually improves the learning model through interaction with the central server. If the FL is applied to blockchain network, it can get many advantages such as security, integrity and efficient incentive system. However, it can cause network overload because of limited communication bandwidth and the participation of a huge number of users. Furthermore, learning speed slow down because of blockchain network. One of the ways to minimize the network load and improve the learning speed are for the model to converge rapidly and stably with target learning accuracy. In this paper, we propose blockchain based federated learning scenario. We will explain of some benefits for blockchain and some methods for improving learning performance. We consider two types of weights to choose the subset of clients for updating the global model. First, we consider the weight based on local learning accuracy of each client. Second, we consider the weight based on participation frequency of each client. We choose two key performance indicators, learning speed and standard deviation, to compare the performance of our proposed scheme with existing schemes. The simulation results show that our proposed scheme achieves higher stability along with fast convergence time for targeted accuracy compared to others.
収録刊行物
-
- IEICE Proceeding Series
-
IEICE Proceeding Series 56 P3-1-, 2019-09-18
The Institute of Electronics, Information and Communication Engineers
- Tweet
詳細情報 詳細情報について
-
- CRID
- 1390848250127256192
-
- NII論文ID
- 230000011552
-
- ISSN
- 21885079
-
- 本文言語コード
- en
-
- データソース種別
-
- JaLC
- CiNii Articles
-
- 抄録ライセンスフラグ
- 使用不可