File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1109/TWC.2020.3015671
- Scopus: eid_2-s2.0-85096408811
- WOS: WOS:000589218700049
- Find via
Supplementary
- Citations:
- Appears in Collections:
Article: Scheduling for Cellular Federated Edge Learning with Importance and Channel Awareness
Title | Scheduling for Cellular Federated Edge Learning with Importance and Channel Awareness |
---|---|
Authors | |
Keywords | Scheduling Convergence Servers Wireless communication Processor scheduling |
Issue Date | 2020 |
Publisher | Institute of Electrical and Electronics Engineers. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=7693 |
Citation | IEEE Transactions on Wireless Communications, 2020, v. 19 n. 11, p. 7690-7703 How to Cite? |
Abstract | In cellular federated edge learning (FEEL), multiple edge devices holding local data jointly train a neural network by communicating learning updates with an access point without exchanging their data samples. With very limited communication resources, it is beneficial to schedule the most informative local learning updates. This paper focuses on FEEL with gradient averaging over participating devices in each round of communication. A novel scheduling policy is proposed to exploit both diversity in multiuser channels and diversity in the “importance” of the edge devices' learning updates. First, a new probabilistic scheduling framework is developed to yield unbiased update aggregation in FEEL. The importance of a local learning update is measured by its gradient divergence. If one edge device is scheduled in each communication round, the scheduling policy is derived in closed form to achieve the optimal trade-off between channel quality and update importance. The probabilistic scheduling framework is then extended to allow scheduling multiple edge devices in each communication round. Numerical results obtained using popular models and learning datasets demonstrate that the proposed scheduling policy can achieve faster model convergence and higher learning accuracy than conventional scheduling policies that only exploit a single type of diversity. |
Persistent Identifier | http://hdl.handle.net/10722/295785 |
ISSN | 2023 Impact Factor: 8.9 2023 SCImago Journal Rankings: 5.371 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Ren, J | - |
dc.contributor.author | He, Y | - |
dc.contributor.author | WEN, D | - |
dc.contributor.author | Yu, G | - |
dc.contributor.author | Huang, K | - |
dc.contributor.author | Guo, D | - |
dc.date.accessioned | 2021-02-08T08:13:59Z | - |
dc.date.available | 2021-02-08T08:13:59Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | IEEE Transactions on Wireless Communications, 2020, v. 19 n. 11, p. 7690-7703 | - |
dc.identifier.issn | 1536-1276 | - |
dc.identifier.uri | http://hdl.handle.net/10722/295785 | - |
dc.description.abstract | In cellular federated edge learning (FEEL), multiple edge devices holding local data jointly train a neural network by communicating learning updates with an access point without exchanging their data samples. With very limited communication resources, it is beneficial to schedule the most informative local learning updates. This paper focuses on FEEL with gradient averaging over participating devices in each round of communication. A novel scheduling policy is proposed to exploit both diversity in multiuser channels and diversity in the “importance” of the edge devices' learning updates. First, a new probabilistic scheduling framework is developed to yield unbiased update aggregation in FEEL. The importance of a local learning update is measured by its gradient divergence. If one edge device is scheduled in each communication round, the scheduling policy is derived in closed form to achieve the optimal trade-off between channel quality and update importance. The probabilistic scheduling framework is then extended to allow scheduling multiple edge devices in each communication round. Numerical results obtained using popular models and learning datasets demonstrate that the proposed scheduling policy can achieve faster model convergence and higher learning accuracy than conventional scheduling policies that only exploit a single type of diversity. | - |
dc.language | eng | - |
dc.publisher | Institute of Electrical and Electronics Engineers. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/RecentIssue.jsp?punumber=7693 | - |
dc.relation.ispartof | IEEE Transactions on Wireless Communications | - |
dc.rights | IEEE Transactions on Wireless Communications. Copyright © Institute of Electrical and Electronics Engineers. | - |
dc.rights | ©20xx IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. | - |
dc.subject | Scheduling | - |
dc.subject | Convergence | - |
dc.subject | Servers | - |
dc.subject | Wireless communication | - |
dc.subject | Processor scheduling | - |
dc.title | Scheduling for Cellular Federated Edge Learning with Importance and Channel Awareness | - |
dc.type | Article | - |
dc.identifier.email | Huang, K: huangkb@eee.hku.hk | - |
dc.identifier.authority | Huang, K=rp01875 | - |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1109/TWC.2020.3015671 | - |
dc.identifier.scopus | eid_2-s2.0-85096408811 | - |
dc.identifier.hkuros | 321245 | - |
dc.identifier.volume | 19 | - |
dc.identifier.issue | 11 | - |
dc.identifier.spage | 7690 | - |
dc.identifier.epage | 7703 | - |
dc.identifier.isi | WOS:000589218700049 | - |
dc.publisher.place | United States | - |