TY - GEN
T1 - Federated model search via reinforcement learning
AU - Yao, Dixi
AU - Wang, Lingdong
AU - Xu, Jiayu
AU - Xiang, Liyao
AU - Shao, Shuo
AU - Chen, Yingqi
AU - Tong, Yanjun
N1 - Publisher Copyright:
© 2021 IEEE.
PY - 2021/7
Y1 - 2021/7
N2 - Federated Learning (FL) framework enables training over distributed datasets while keeping the data local. However, it is difficult to customize a model fitting for all unknown local data. A pre-determined model is most likely to lead to slow convergence or low accuracy, especially when the distributed data is non-i.i.d.. To resolve the issue, we propose a model searching method in the federated learning scenario, and the method automatically searches a model structure fitting for the unseen local data. We novelly design a reinforcement learning-based framework that samples and distributes sub-models to the participants and updates its model selection policy by maximizing the reward. In practice, the model search algorithm takes a long time to converge, and hence we adaptively assign sub-models to participants according to the transmission condition. We further propose delay-compensated synchronization to mitigate loss over late updates to facilitate convergence. Extensive experiments show that our federated model search algorithm produces highly accurate models efficiently, particularly on non-i.i.d. data.
AB - Federated Learning (FL) framework enables training over distributed datasets while keeping the data local. However, it is difficult to customize a model fitting for all unknown local data. A pre-determined model is most likely to lead to slow convergence or low accuracy, especially when the distributed data is non-i.i.d.. To resolve the issue, we propose a model searching method in the federated learning scenario, and the method automatically searches a model structure fitting for the unseen local data. We novelly design a reinforcement learning-based framework that samples and distributes sub-models to the participants and updates its model selection policy by maximizing the reward. In practice, the model search algorithm takes a long time to converge, and hence we adaptively assign sub-models to participants according to the transmission condition. We further propose delay-compensated synchronization to mitigate loss over late updates to facilitate convergence. Extensive experiments show that our federated model search algorithm produces highly accurate models efficiently, particularly on non-i.i.d. data.
KW - Edge Cloud Computing
KW - Neural Architecture Search
KW - Reinforcement Learning
UR - https://www.scopus.com/pages/publications/85117158095
U2 - 10.1109/ICDCS51616.2021.00084
DO - 10.1109/ICDCS51616.2021.00084
M3 - 会议稿件
AN - SCOPUS:85117158095
T3 - Proceedings - International Conference on Distributed Computing Systems
SP - 830
EP - 840
BT - Proceedings - 2021 IEEE 41st International Conference on Distributed Computing Systems, ICDCS 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 41st IEEE International Conference on Distributed Computing Systems, ICDCS 2021
Y2 - 7 July 2021 through 10 July 2021
ER -