DAGC: Data-Volume-Aware Adaptive Sparsification Gradient Compression for Distributed Machine Learning in Mobile Computing ...
Distributed machine learning (DML) in mobile environments faces significant communication bottlenecks. Gradient compression has emerged as an effective solution to this issue, offering substantial benefits in environments with limited bandwidth and metered data. Yet, they encounter severe performanc...
Main Authors: | , , , , , , |
---|---|
Format: | Report |
Language: | unknown |
Published: |
arXiv
2023
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.2311.07324 https://arxiv.org/abs/2311.07324 |
id |
ftdatacite:10.48550/arxiv.2311.07324 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.2311.07324 2023-12-31T10:06:17+01:00 DAGC: Data-Volume-Aware Adaptive Sparsification Gradient Compression for Distributed Machine Learning in Mobile Computing ... Lu, Rongwei Jiang, Yutong Mao, Yinan Tang, Chen Chen, Bin Cui, Laizhong Wang, Zhi 2023 https://dx.doi.org/10.48550/arxiv.2311.07324 https://arxiv.org/abs/2311.07324 unknown arXiv Creative Commons Attribution Non Commercial Share Alike 4.0 International https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode cc-by-nc-sa-4.0 Machine Learning cs.LG FOS Computer and information sciences CreativeWork Preprint article Article 2023 ftdatacite https://doi.org/10.48550/arxiv.2311.07324 2023-12-01T11:12:53Z Distributed machine learning (DML) in mobile environments faces significant communication bottlenecks. Gradient compression has emerged as an effective solution to this issue, offering substantial benefits in environments with limited bandwidth and metered data. Yet, they encounter severe performance drop in non-IID environments due to a one-size-fits-all compression approach, which does not account for the varying data volumes across workers. Assigning varying compression ratios to workers with distinct data distributions and volumes is thus a promising solution. This study introduces an analysis of distributed SGD with non-uniform compression, which reveals that the convergence rate (indicative of the iterations needed to achieve a certain accuracy) is influenced by compression ratios applied to workers with differing volumes. Accordingly, we frame relative compression ratio assignment as an $n$-variables chi-square nonlinear optimization problem, constrained by a fixed and limited communication budget. We ... Report DML DataCite Metadata Store (German National Library of Science and Technology) |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
unknown |
topic |
Machine Learning cs.LG FOS Computer and information sciences |
spellingShingle |
Machine Learning cs.LG FOS Computer and information sciences Lu, Rongwei Jiang, Yutong Mao, Yinan Tang, Chen Chen, Bin Cui, Laizhong Wang, Zhi DAGC: Data-Volume-Aware Adaptive Sparsification Gradient Compression for Distributed Machine Learning in Mobile Computing ... |
topic_facet |
Machine Learning cs.LG FOS Computer and information sciences |
description |
Distributed machine learning (DML) in mobile environments faces significant communication bottlenecks. Gradient compression has emerged as an effective solution to this issue, offering substantial benefits in environments with limited bandwidth and metered data. Yet, they encounter severe performance drop in non-IID environments due to a one-size-fits-all compression approach, which does not account for the varying data volumes across workers. Assigning varying compression ratios to workers with distinct data distributions and volumes is thus a promising solution. This study introduces an analysis of distributed SGD with non-uniform compression, which reveals that the convergence rate (indicative of the iterations needed to achieve a certain accuracy) is influenced by compression ratios applied to workers with differing volumes. Accordingly, we frame relative compression ratio assignment as an $n$-variables chi-square nonlinear optimization problem, constrained by a fixed and limited communication budget. We ... |
format |
Report |
author |
Lu, Rongwei Jiang, Yutong Mao, Yinan Tang, Chen Chen, Bin Cui, Laizhong Wang, Zhi |
author_facet |
Lu, Rongwei Jiang, Yutong Mao, Yinan Tang, Chen Chen, Bin Cui, Laizhong Wang, Zhi |
author_sort |
Lu, Rongwei |
title |
DAGC: Data-Volume-Aware Adaptive Sparsification Gradient Compression for Distributed Machine Learning in Mobile Computing ... |
title_short |
DAGC: Data-Volume-Aware Adaptive Sparsification Gradient Compression for Distributed Machine Learning in Mobile Computing ... |
title_full |
DAGC: Data-Volume-Aware Adaptive Sparsification Gradient Compression for Distributed Machine Learning in Mobile Computing ... |
title_fullStr |
DAGC: Data-Volume-Aware Adaptive Sparsification Gradient Compression for Distributed Machine Learning in Mobile Computing ... |
title_full_unstemmed |
DAGC: Data-Volume-Aware Adaptive Sparsification Gradient Compression for Distributed Machine Learning in Mobile Computing ... |
title_sort |
dagc: data-volume-aware adaptive sparsification gradient compression for distributed machine learning in mobile computing ... |
publisher |
arXiv |
publishDate |
2023 |
url |
https://dx.doi.org/10.48550/arxiv.2311.07324 https://arxiv.org/abs/2311.07324 |
genre |
DML |
genre_facet |
DML |
op_rights |
Creative Commons Attribution Non Commercial Share Alike 4.0 International https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode cc-by-nc-sa-4.0 |
op_doi |
https://doi.org/10.48550/arxiv.2311.07324 |
_version_ |
1786838271947440128 |