Synchronization-Aware NAS for an Efficient Collaborative Inference on Mobile Platforms
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kang, Beom Woo | - |
dc.contributor.author | Wohn, Junho | - |
dc.contributor.author | Lee, Seongju | - |
dc.contributor.author | Park, Sunghyun | - |
dc.contributor.author | Noh, Yung-Kyun | - |
dc.contributor.author | Park, Yongjun | - |
dc.date.accessioned | 2023-08-01T06:34:52Z | - |
dc.date.available | 2023-08-01T06:34:52Z | - |
dc.date.issued | 2023-06 | - |
dc.identifier.uri | https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/188399 | - |
dc.description.abstract | Previous neural architecture search (NAS) approaches for mobile platforms have achieved great success in designing a slim-but-accurate neural network that is generally well-matched to a single computing unit such as a CPU or GPU. However, as recent mobile devices consist of multiple heterogeneous computing units, the next main challenge is to maximize both accuracy and efficiency by fully utilizing multiple available resources. We propose an ensemble-like approach with intermediate feature aggregations, namely synchronizations, for active collaboration between individual models on a mobile device. A main challenge is to determine the optimal synchronization strategies for achieving both performance and efficiency. To this end, we propose SyncNAS to automate the exploration of synchronization strategies for collaborative neural architectures that maximize utilization of heterogeneous computing units on a target device. We introduce a novel search space for synchronization strategy and apply Monte Carlo tree search (MCTS) algorithm to improve the sampling efficiency and reduce the search cost. On ImageNet, our collaborative model based on MobileNetV2 achieves 2.7% top-1 accuracy improvement within the baseline latency budget. Under the reduced target latency down to half, our model maintains higher accuracy than its baseline model, owing to the enhanced utilization and collaboration. As an impact of MCTS, SyncNAS reduces its search cost by up to 21× in searching for the optimal strategy. | - |
dc.format.extent | 13 | - |
dc.language | 영어 | - |
dc.language.iso | ENG | - |
dc.publisher | Association for Computing Machinery | - |
dc.title | Synchronization-Aware NAS for an Efficient Collaborative Inference on Mobile Platforms | - |
dc.type | Article | - |
dc.publisher.location | 국제연합 | - |
dc.identifier.doi | 10.1145/3589610.3596284 | - |
dc.identifier.scopusid | 2-s2.0-85164276609 | - |
dc.identifier.wosid | 001117978800003 | - |
dc.identifier.bibliographicCitation | Proceedings of the ACM SIGPLAN Conference on Languages, Compilers, and Tools for Embedded Systems (LCTES), pp 13 - 25 | - |
dc.citation.title | Proceedings of the ACM SIGPLAN Conference on Languages, Compilers, and Tools for Embedded Systems (LCTES) | - |
dc.citation.startPage | 13 | - |
dc.citation.endPage | 25 | - |
dc.type.docType | Proceedings Paper | - |
dc.description.isOpenAccess | N | - |
dc.description.journalRegisteredClass | scopus | - |
dc.relation.journalResearchArea | Computer Science | - |
dc.relation.journalWebOfScienceCategory | Computer Science, Hardware & Architecture | - |
dc.relation.journalWebOfScienceCategory | Computer Science, Software Engineering | - |
dc.relation.journalWebOfScienceCategory | Computer Science, Theory & Methods | - |
dc.subject.keywordPlus | Budget control | - |
dc.subject.keywordPlus | Cost reduction | - |
dc.subject.keywordPlus | Drilling platforms | - |
dc.subject.keywordPlus | Efficiency | - |
dc.subject.keywordPlus | Image enhancement | - |
dc.subject.keywordPlus | Network architecture | - |
dc.subject.keywordPlus | Trees (mathematics) | - |
dc.subject.keywordPlus | Synchronization | - |
dc.subject.keywordPlus | Computing units | - |
dc.subject.keywordPlus | Heterogeneous computing | - |
dc.subject.keywordPlus | Mobile platform | - |
dc.subject.keywordPlus | Model parallelism | - |
dc.subject.keywordPlus | Neural architecture search | - |
dc.subject.keywordPlus | Neural architectures | - |
dc.subject.keywordPlus | Neural-networks | - |
dc.subject.keywordPlus | On-device ML | - |
dc.subject.keywordPlus | Search costs | - |
dc.subject.keywordPlus | Synchronization strategies | - |
dc.subject.keywordAuthor | Model Parallelism | - |
dc.subject.keywordAuthor | Neural Architecture Search | - |
dc.subject.keywordAuthor | On-Device ML | - |
dc.identifier.url | https://dl.acm.org/doi/10.1145/3589610.3596284 | - |
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.
222, Wangsimni-ro, Seongdong-gu, Seoul, 04763, Korea+82-2-2220-1365
COPYRIGHT © 2021 HANYANG UNIVERSITY.
Certain data included herein are derived from the © Web of Science of Clarivate Analytics. All rights reserved.
You may not copy or re-distribute this material in whole or in part without the prior written consent of Clarivate Analytics.