講演名 2023-08-31
Performance Analysis of On-device Hierarchical Federated Learning Frameworks
杜 兆陽(電通大), 策力 木格(電通大), 吉永 努(電通大),
PDFダウンロードページ PDFダウンロードページへ
抄録(和) In the rapidly advancing field of artificial intelligence and deep learning, centralized architectures exhibit inherent limitations such as high latency and computational overhead. This study introduces a Hierarchical Federated Learning (HFL) framework to overcome these challenges by decentralizing the learning process. Employing intermediary edge servers, HFL efficiently distributes computational and communication loads, as demonstrated through experiments with Raspberry Pi devices and the Cifar10 dataset. The results reveal significant improvements in model accuracy and convergence speed over traditional cloud-based Federated Learning (FL), showcasing the potential of HFL to enhance FL performance with acceptable communication costs. This innovation marks a step towards more resilient and adaptive learning systems in areas such as Intelligent Transportation Systems (ITS) and Internet of Things (IoT) devices.
抄録(英) In the rapidly advancing field of artificial intelligence and deep learning, centralized architectures exhibit inherent limitations such as high latency and computational overhead. This study introduces a Hierarchical Federated Learning (HFL) framework to overcome these challenges by decentralizing the learning process. Employing intermediary edge servers, HFL efficiently distributes computational and communication loads, as demonstrated through experiments with Raspberry Pi devices and the Cifar10 dataset. The results reveal significant improvements in model accuracy and convergence speed over traditional cloud-based Federated Learning (FL), showcasing the potential of HFL to enhance FL performance with acceptable communication costs. This innovation marks a step towards more resilient and adaptive learning systems in areas such as Intelligent Transportation Systems (ITS) and Internet of Things (IoT) devices.
キーワード(和) Hierarchical Federated Learning / Edge Computing / IoT / Intelligent Transportation Systems
キーワード(英) Hierarchical Federated Learning / Edge Computing / IoT / Intelligent Transportation Systems
資料番号 CQ2023-28
発行日 2023-08-24 (CQ)

研究会情報
研究会 CQ
開催期間 2023/8/31(から2日開催)
開催地(和) 天神岬スポーツ公園
開催地(英) Tenjin-Misaki Sports Park
テーマ(和) 無線通信品質、6G、IoT、無線伝送、リソース制御、クロスレイヤー技術、一般
テーマ(英) Wireless Communications Quality, 6G, IoT, Resource Management, Wireless Transmission, Cross layer Technologies, etc.
委員長氏名(和) 平栗 健史(日本工大)
委員長氏名(英) Takefumi Hiraguri(Nippon Inst. of Tech.)
副委員長氏名(和) 松田 崇弘(都立大) / 長谷川 剛(東北大) / 新井田 統(KDDI総合研究所)
副委員長氏名(英) Takahiro Matsuda(Tokyo Metropolitan Univ.) / Gou Hasegawa(Tohoku Univ.) / Sumaru Niida(KDDI Research)
幹事氏名(和) 恵木 則次(NTT) / 菅沼 睦(多摩大) / 津川 翔(筑波大)
幹事氏名(英) Noritsugu Egi(NTT) / Mutsumi Suganuma(Tama Univ.) / Shou Tsugawa(Tsukuba Univ.)
幹事補佐氏名(和) 中村 遼(福岡大) / 中平 俊朗(NTT) / 塚常 健太(岡山理科大)
幹事補佐氏名(英) Ryo Nakamura(Fukuoka Univ.) / Toshiro Nakahira(NTT) / Kenta Tsukatsune(Okayama Univ. of Science)

講演論文情報詳細
申込み研究会 Technical Committee on Communication Quality
本文の言語 ENG
タイトル(和)
サブタイトル(和)
タイトル(英) Performance Analysis of On-device Hierarchical Federated Learning Frameworks
サブタイトル(和)
キーワード(1)(和/英) Hierarchical Federated Learning / Hierarchical Federated Learning
キーワード(2)(和/英) Edge Computing / Edge Computing
キーワード(3)(和/英) IoT / IoT
キーワード(4)(和/英) Intelligent Transportation Systems / Intelligent Transportation Systems
第 1 著者 氏名(和/英) 杜 兆陽 / Zhaoyang Du
第 1 著者 所属(和/英) 電気通信大学(略称:電通大)
The University of Electro-Communications(略称:UEC)
第 2 著者 氏名(和/英) 策力 木格 / Celimuge Wu
第 2 著者 所属(和/英) 電気通信大学(略称:電通大)
The University of Electro-Communications(略称:UEC)
第 3 著者 氏名(和/英) 吉永 努 / Tsutomu Yoshinaga
第 3 著者 所属(和/英) 電気通信大学(略称:電通大)
The University of Electro-Communications(略称:UEC)
発表年月日 2023-08-31
資料番号 CQ2023-28
巻番号(vol) vol.123
号番号(no) CQ-174
ページ範囲 pp.14-19(CQ),
ページ数 6
発行日 2023-08-24 (CQ)