PyramidFL: A fine-grained client selection framework for effiicient federated learning

Shibo Wang; Shusen Yang; Hailiang Li; Xiaodan Zhang; Chen Zhou; Chenren Xu; Feng Qian; Nanbin Wang; Zongben Xu

doi:10.1145/3495243.3517018

PyramidFL: A fine-grained client selection framework for effiicient federated learning

Shibo Wang, Shusen Yang, Hailiang Li, Xiaodan Zhang, Chen Zhou, Chenren Xu, Feng Qian, Nanbin Wang, Zongben Xu

Computer Science and Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

4 Scopus citations

Abstract

Mobile 360° video streaming has grown significantly in popularity but the quality of experience (QoE) suffers from insufficient wireless network bandwidth. The state-of-The-Art solutions are limited by the temporal correlation assumption. Recent studies are aware of the potential of saliency to further QoE improvement, but several fundamental challenges about saliency judgment, saliency acquirement, and quality adaptation are still not fully addressed. To solve these challenges, we present SalientVR, a saliency-driven mobile 360° video streaming system integrated with gaze information. We design (i) a precise gaze-driven saliency judging criterion for mobile VR viewers, (ii) two pragmatic gaze-driven, tile-level saliency acquiring methods based on cross-user similarity and a specific content-Aware deep neural network respectively, and (iii) a lightweight saliency-Aware quality adaptation algorithm with a motion-Assisted online correction, which is robust to wireless bandwidth vagaries and saliency bias. Moreover, we contribute a gaze-Annotated dataset and a gaze-driven quality assessment metric for 360° videos. By extensive prototype evaluations (based on dataset tests and user studies), compared to alternatives, SalientVR significantly enhances the video quality and reduces the rebuffering ratio over 4G/LTE network emulations and in the wild, which achieves a 43.68% QoE improvement.

Original language	English (US)
Title of host publication	ACM MobiCom 2022 - Proceedings of the 2022 28th Annual International Conference on Mobile Computing and Networking
Publisher	Association for Computing Machinery
Pages	542-555
Number of pages	14
ISBN (Electronic)	9781450391818
DOIs	https://doi.org/10.1145/3495243.3517018
State	Published - Oct 14 2022
Event	28th ACM Annual International Conference on Mobile Computing and Networking, MobiCom 2022 - Sydney, Australia Duration: Oct 17 2202 → Oct 21 2202

Publication series

Name	Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM

Conference

Conference	28th ACM Annual International Conference on Mobile Computing and Networking, MobiCom 2022
Country/Territory	Australia
City	Sydney
Period	10/17/02 → 10/21/02

Bibliographical note

Funding Information:
We would like to thank the anonymous reviewers and shepherd for their thoughtful suggestions. This work was supported by the National Key Research and Development Program of China under Grant 2020YFA0713900; the National Natural Science Foundation of China under Grant 61772410, Grant 61802298, Grant 62172329, Grant U1811461, Grant 11690011, Grant U21A6005, Grant 62022005, and Grant 62061146001; and the China Postdoctoral Science Foundation under Grant 2020T130513 and Grant 2019M663726.

Publisher Copyright:
© 2022 ACM.

Access

10.1145/3495243.3517018

OpenUrl availability

Full text

Cite this

Wang, S., Yang, S., Li, H., Zhang, X., Zhou, C., Xu, C., Qian, F., Wang, N., & Xu, Z. (2022). PyramidFL: A fine-grained client selection framework for effiicient federated learning. In ACM MobiCom 2022 - Proceedings of the 2022 28th Annual International Conference on Mobile Computing and Networking (pp. 542-555). (Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM). Association for Computing Machinery. https://doi.org/10.1145/3495243.3517018

PyramidFL: A fine-grained client selection framework for effiicient federated learning. / Wang, Shibo; Yang, Shusen; Li, Hailiang et al.
ACM MobiCom 2022 - Proceedings of the 2022 28th Annual International Conference on Mobile Computing and Networking. Association for Computing Machinery, 2022. p. 542-555 (Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Wang, S, Yang, S, Li, H, Zhang, X, Zhou, C, Xu, C, Qian, F, Wang, N & Xu, Z 2022, PyramidFL: A fine-grained client selection framework for effiicient federated learning. in ACM MobiCom 2022 - Proceedings of the 2022 28th Annual International Conference on Mobile Computing and Networking. Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM, Association for Computing Machinery, pp. 542-555, 28th ACM Annual International Conference on Mobile Computing and Networking, MobiCom 2022, Sydney, Australia, 10/17/02. https://doi.org/10.1145/3495243.3517018

Wang S, Yang S, Li H, Zhang X, Zhou C, Xu C et al. PyramidFL: A fine-grained client selection framework for effiicient federated learning. In ACM MobiCom 2022 - Proceedings of the 2022 28th Annual International Conference on Mobile Computing and Networking. Association for Computing Machinery. 2022. p. 542-555. (Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM). doi: 10.1145/3495243.3517018

Wang, Shibo ; Yang, Shusen ; Li, Hailiang et al. / PyramidFL : A fine-grained client selection framework for effiicient federated learning. ACM MobiCom 2022 - Proceedings of the 2022 28th Annual International Conference on Mobile Computing and Networking. Association for Computing Machinery, 2022. pp. 542-555 (Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM).

@inproceedings{85e883699a1440d49b81ed227b2f9f1a,

title = "PyramidFL: A fine-grained client selection framework for effiicient federated learning",

abstract = "Mobile 360° video streaming has grown significantly in popularity but the quality of experience (QoE) suffers from insufficient wireless network bandwidth. The state-of-The-Art solutions are limited by the temporal correlation assumption. Recent studies are aware of the potential of saliency to further QoE improvement, but several fundamental challenges about saliency judgment, saliency acquirement, and quality adaptation are still not fully addressed. To solve these challenges, we present SalientVR, a saliency-driven mobile 360° video streaming system integrated with gaze information. We design (i) a precise gaze-driven saliency judging criterion for mobile VR viewers, (ii) two pragmatic gaze-driven, tile-level saliency acquiring methods based on cross-user similarity and a specific content-Aware deep neural network respectively, and (iii) a lightweight saliency-Aware quality adaptation algorithm with a motion-Assisted online correction, which is robust to wireless bandwidth vagaries and saliency bias. Moreover, we contribute a gaze-Annotated dataset and a gaze-driven quality assessment metric for 360° videos. By extensive prototype evaluations (based on dataset tests and user studies), compared to alternatives, SalientVR significantly enhances the video quality and reduces the rebuffering ratio over 4G/LTE network emulations and in the wild, which achieves a 43.68% QoE improvement.",

author = "Shibo Wang and Shusen Yang and Hailiang Li and Xiaodan Zhang and Chen Zhou and Chenren Xu and Feng Qian and Nanbin Wang and Zongben Xu",

note = "Funding Information: We would like to thank the anonymous reviewers and shepherd for their thoughtful suggestions. This work was supported by the National Key Research and Development Program of China under Grant 2020YFA0713900; the National Natural Science Foundation of China under Grant 61772410, Grant 61802298, Grant 62172329, Grant U1811461, Grant 11690011, Grant U21A6005, Grant 62022005, and Grant 62061146001; and the China Postdoctoral Science Foundation under Grant 2020T130513 and Grant 2019M663726. Publisher Copyright: {\textcopyright} 2022 ACM.; 28th ACM Annual International Conference on Mobile Computing and Networking, MobiCom 2022 ; Conference date: 17-10-2202 Through 21-10-2202",

year = "2022",

month = oct,

day = "14",

doi = "10.1145/3495243.3517018",

language = "English (US)",

series = "Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM",

publisher = "Association for Computing Machinery",

pages = "542--555",

booktitle = "ACM MobiCom 2022 - Proceedings of the 2022 28th Annual International Conference on Mobile Computing and Networking",

}

TY - GEN

T1 - PyramidFL

T2 - 28th ACM Annual International Conference on Mobile Computing and Networking, MobiCom 2022

AU - Wang, Shibo

AU - Yang, Shusen

AU - Li, Hailiang

AU - Zhang, Xiaodan

AU - Zhou, Chen

AU - Xu, Chenren

AU - Qian, Feng

AU - Wang, Nanbin

AU - Xu, Zongben

N1 - Funding Information: We would like to thank the anonymous reviewers and shepherd for their thoughtful suggestions. This work was supported by the National Key Research and Development Program of China under Grant 2020YFA0713900; the National Natural Science Foundation of China under Grant 61772410, Grant 61802298, Grant 62172329, Grant U1811461, Grant 11690011, Grant U21A6005, Grant 62022005, and Grant 62061146001; and the China Postdoctoral Science Foundation under Grant 2020T130513 and Grant 2019M663726. Publisher Copyright: © 2022 ACM.

PY - 2022/10/14

Y1 - 2022/10/14

N2 - Mobile 360° video streaming has grown significantly in popularity but the quality of experience (QoE) suffers from insufficient wireless network bandwidth. The state-of-The-Art solutions are limited by the temporal correlation assumption. Recent studies are aware of the potential of saliency to further QoE improvement, but several fundamental challenges about saliency judgment, saliency acquirement, and quality adaptation are still not fully addressed. To solve these challenges, we present SalientVR, a saliency-driven mobile 360° video streaming system integrated with gaze information. We design (i) a precise gaze-driven saliency judging criterion for mobile VR viewers, (ii) two pragmatic gaze-driven, tile-level saliency acquiring methods based on cross-user similarity and a specific content-Aware deep neural network respectively, and (iii) a lightweight saliency-Aware quality adaptation algorithm with a motion-Assisted online correction, which is robust to wireless bandwidth vagaries and saliency bias. Moreover, we contribute a gaze-Annotated dataset and a gaze-driven quality assessment metric for 360° videos. By extensive prototype evaluations (based on dataset tests and user studies), compared to alternatives, SalientVR significantly enhances the video quality and reduces the rebuffering ratio over 4G/LTE network emulations and in the wild, which achieves a 43.68% QoE improvement.

AB - Mobile 360° video streaming has grown significantly in popularity but the quality of experience (QoE) suffers from insufficient wireless network bandwidth. The state-of-The-Art solutions are limited by the temporal correlation assumption. Recent studies are aware of the potential of saliency to further QoE improvement, but several fundamental challenges about saliency judgment, saliency acquirement, and quality adaptation are still not fully addressed. To solve these challenges, we present SalientVR, a saliency-driven mobile 360° video streaming system integrated with gaze information. We design (i) a precise gaze-driven saliency judging criterion for mobile VR viewers, (ii) two pragmatic gaze-driven, tile-level saliency acquiring methods based on cross-user similarity and a specific content-Aware deep neural network respectively, and (iii) a lightweight saliency-Aware quality adaptation algorithm with a motion-Assisted online correction, which is robust to wireless bandwidth vagaries and saliency bias. Moreover, we contribute a gaze-Annotated dataset and a gaze-driven quality assessment metric for 360° videos. By extensive prototype evaluations (based on dataset tests and user studies), compared to alternatives, SalientVR significantly enhances the video quality and reduces the rebuffering ratio over 4G/LTE network emulations and in the wild, which achieves a 43.68% QoE improvement.

UR - http://www.scopus.com/inward/record.url?scp=85140879395&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85140879395&partnerID=8YFLogxK

U2 - 10.1145/3495243.3517018

DO - 10.1145/3495243.3517018

M3 - Conference contribution

AN - SCOPUS:85140879395

T3 - Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM

SP - 542

EP - 555

BT - ACM MobiCom 2022 - Proceedings of the 2022 28th Annual International Conference on Mobile Computing and Networking

PB - Association for Computing Machinery

Y2 - 17 October 2202 through 21 October 2202

ER -

PyramidFL: A fine-grained client selection framework for effiicient federated learning

Abstract

Publication series

Conference

Bibliographical note

Access

OpenUrl availability

Other files and links

Fingerprint

Cite this