File Download
There are no files associated with this item.
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1109/GLOBECOM42002.2020.9347992
- Scopus: eid_2-s2.0-85100873814
- WOS: WOS:000668970503150
- Find via
Supplementary
- Citations:
- Appears in Collections:
Conference Paper: Accelerating Partitioned Edge Learning via Joint Parameter-and-Bandwidth Allocation
Title | Accelerating Partitioned Edge Learning via Joint Parameter-and-Bandwidth Allocation |
---|---|
Authors | |
Keywords | Training Wireless networks Stochastic processes Resource management Channel allocation |
Issue Date | 2020 |
Publisher | Institute of Electrical and Electronics Engineers. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1000308 |
Citation | Proceedings of GLOBECOM 2020 - 2020 IEEE Global Communications Conference, Virtual Conference, Taipei, Taiwan, 7-11 December 2020, p. 1-6 How to Cite? |
Abstract | In this paper, we consider the framework of partitioned edge learning for iteratively training a large-scale model using many resource-constrained devices (called workers). To this end, in each iteration, the model is dynamically partitioned into parametric blocks, which are downloaded to worker groups for updating using their local data. Then, the local updates are uploaded to and cascaded by the server for updating a global model. To reduce resource usage by minimizing the total learning-and-communication latency, this work focuses on the novel joint design of parameter (computation load) and bandwidth allocation (for downloading and uploading). Two design approaches are adopted. First, a practical sequential approach, called partially integrated parameter-and-bandwidth allocation (PABA), yields one scheme, namely parameter aware bandwidth allocation. It allocates the largest bandwidth to the slowest worker. Second, PABA are jointly optimized. Despite its being a nonconvex problem, an efficient and optimal solution algorithm is derived by intelligently nesting a bisection search and solving a convex problem. Experimental results using real data demonstrate that integrating PABA can substantially improve the performance of partitioned edge learning in terms of latency (by e.g., 46%) and accuracy (by e.g., 4%). |
Persistent Identifier | http://hdl.handle.net/10722/295865 |
ISSN | |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Wen, D | - |
dc.contributor.author | Bennis, M | - |
dc.contributor.author | Huang, K | - |
dc.date.accessioned | 2021-02-08T08:15:08Z | - |
dc.date.available | 2021-02-08T08:15:08Z | - |
dc.date.issued | 2020 | - |
dc.identifier.citation | Proceedings of GLOBECOM 2020 - 2020 IEEE Global Communications Conference, Virtual Conference, Taipei, Taiwan, 7-11 December 2020, p. 1-6 | - |
dc.identifier.issn | 2334-0983 | - |
dc.identifier.uri | http://hdl.handle.net/10722/295865 | - |
dc.description.abstract | In this paper, we consider the framework of partitioned edge learning for iteratively training a large-scale model using many resource-constrained devices (called workers). To this end, in each iteration, the model is dynamically partitioned into parametric blocks, which are downloaded to worker groups for updating using their local data. Then, the local updates are uploaded to and cascaded by the server for updating a global model. To reduce resource usage by minimizing the total learning-and-communication latency, this work focuses on the novel joint design of parameter (computation load) and bandwidth allocation (for downloading and uploading). Two design approaches are adopted. First, a practical sequential approach, called partially integrated parameter-and-bandwidth allocation (PABA), yields one scheme, namely parameter aware bandwidth allocation. It allocates the largest bandwidth to the slowest worker. Second, PABA are jointly optimized. Despite its being a nonconvex problem, an efficient and optimal solution algorithm is derived by intelligently nesting a bisection search and solving a convex problem. Experimental results using real data demonstrate that integrating PABA can substantially improve the performance of partitioned edge learning in terms of latency (by e.g., 46%) and accuracy (by e.g., 4%). | - |
dc.language | eng | - |
dc.publisher | Institute of Electrical and Electronics Engineers. The Journal's web site is located at http://ieeexplore.ieee.org/xpl/conhome.jsp?punumber=1000308 | - |
dc.relation.ispartof | IEEE Global Communications Conference (GLOBECOM) | - |
dc.rights | IEEE Global Communications Conference (GLOBECOM). Copyright © Institute of Electrical and Electronics Engineers. | - |
dc.rights | ©2020 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 | Training | - |
dc.subject | Wireless networks | - |
dc.subject | Stochastic processes | - |
dc.subject | Resource management | - |
dc.subject | Channel allocation | - |
dc.title | Accelerating Partitioned Edge Learning via Joint Parameter-and-Bandwidth Allocation | - |
dc.type | Conference_Paper | - |
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/GLOBECOM42002.2020.9347992 | - |
dc.identifier.scopus | eid_2-s2.0-85100873814 | - |
dc.identifier.hkuros | 321259 | - |
dc.identifier.spage | 1 | - |
dc.identifier.epage | 6 | - |
dc.identifier.isi | WOS:000668970503150 | - |
dc.publisher.place | United States | - |
dc.identifier.issnl | 2334-0983 | - |