AdaEmb-Encoder: Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data

Yaobin Qin; David J. Lilja

doi:10.1109/IPCCC50635.2020.9391523

AdaEmb-Encoder: Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data

Yaobin Qin, David J. Lilja

Electrical and Computer Engineering

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

1 Scopus citations

Abstract

The advent of the AI era has made it increasingly important to have an efficient backup system to protect training data from loss. Furthermore, a backup of the training data makes it possible to update or retrain the learned model as more data are collected. However, a huge backup overhead will result if a complete copy of all daily collected training data is always made to backup storage, especially because the data typically contain highly redundant information that makes no contribution to model learning. Deduplication is a common technique in modern backup systems to reduce data redundancy. However, existing deduplication methods are invalid for training data. Hence, this paper proposes a novel deduplication strategy for the training data used for learning in a deep neural network classifier. Experimental results showed that the proposed deduplication strategy achieved 93% backup storage space reduction with only 1.3% loss of classification accuracy.

Original language	English (US)
Title of host publication	2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781728198293
DOIs	https://doi.org/10.1109/IPCCC50635.2020.9391523
State	Published - Nov 6 2020
Event	39th IEEE International Performance Computing and Communications Conference, IPCCC 2020 - Austin, United States Duration: Nov 6 2020 → Nov 8 2020

Publication series

Name	2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020

Conference

Conference	39th IEEE International Performance Computing and Communications Conference, IPCCC 2020
Country/Territory	United States
City	Austin
Period	11/6/20 → 11/8/20

Bibliographical note

Funding Information:
This work was supported in part by the Center for Research in Intelligent Storage (CRIS), which is supported by National Science Foundation grant no. IIP-1439622 and member companies. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF.

Publisher Copyright:
© 2020 IEEE.

Keywords

Backup systems
Deduplication
Deep learning
Training data

Access

10.1109/IPCCC50635.2020.9391523

OpenUrl availability

Full text

Cite this

Qin, Y., & Lilja, D. J. (2020). AdaEmb-Encoder: Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data. In 2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020 Article 9391523 (2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IPCCC50635.2020.9391523

AdaEmb-Encoder: Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data. / Qin, Yaobin; Lilja, David J.
2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020. Institute of Electrical and Electronics Engineers Inc., 2020. 9391523 (2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020).

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

Qin, Y & Lilja, DJ 2020, AdaEmb-Encoder: Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data. in 2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020., 9391523, 2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020, Institute of Electrical and Electronics Engineers Inc., 39th IEEE International Performance Computing and Communications Conference, IPCCC 2020, Austin, United States, 11/6/20. https://doi.org/10.1109/IPCCC50635.2020.9391523

Qin Y, Lilja DJ. AdaEmb-Encoder: Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data. In 2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020. Institute of Electrical and Electronics Engineers Inc. 2020. 9391523. (2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020). doi: 10.1109/IPCCC50635.2020.9391523

Qin, Yaobin ; Lilja, David J. / AdaEmb-Encoder : Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data. 2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020. Institute of Electrical and Electronics Engineers Inc., 2020. (2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020).

@inproceedings{013a21151bac425ea17843a97e3fbbb7,

title = "AdaEmb-Encoder: Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data",

abstract = "The advent of the AI era has made it increasingly important to have an efficient backup system to protect training data from loss. Furthermore, a backup of the training data makes it possible to update or retrain the learned model as more data are collected. However, a huge backup overhead will result if a complete copy of all daily collected training data is always made to backup storage, especially because the data typically contain highly redundant information that makes no contribution to model learning. Deduplication is a common technique in modern backup systems to reduce data redundancy. However, existing deduplication methods are invalid for training data. Hence, this paper proposes a novel deduplication strategy for the training data used for learning in a deep neural network classifier. Experimental results showed that the proposed deduplication strategy achieved 93% backup storage space reduction with only 1.3% loss of classification accuracy.",

keywords = "Backup systems, Deduplication, Deep learning, Training data",

author = "Yaobin Qin and Lilja, {David J.}",

note = "Funding Information: This work was supported in part by the Center for Research in Intelligent Storage (CRIS), which is supported by National Science Foundation grant no. IIP-1439622 and member companies. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF. Publisher Copyright: {\textcopyright} 2020 IEEE.; 39th IEEE International Performance Computing and Communications Conference, IPCCC 2020 ; Conference date: 06-11-2020 Through 08-11-2020",

year = "2020",

month = nov,

day = "6",

doi = "10.1109/IPCCC50635.2020.9391523",

language = "English (US)",

series = "2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020",

}

TY - GEN

T1 - AdaEmb-Encoder

T2 - 39th IEEE International Performance Computing and Communications Conference, IPCCC 2020

AU - Qin, Yaobin

AU - Lilja, David J.

N1 - Funding Information: This work was supported in part by the Center for Research in Intelligent Storage (CRIS), which is supported by National Science Foundation grant no. IIP-1439622 and member companies. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF. Publisher Copyright: © 2020 IEEE.

PY - 2020/11/6

Y1 - 2020/11/6

N2 - The advent of the AI era has made it increasingly important to have an efficient backup system to protect training data from loss. Furthermore, a backup of the training data makes it possible to update or retrain the learned model as more data are collected. However, a huge backup overhead will result if a complete copy of all daily collected training data is always made to backup storage, especially because the data typically contain highly redundant information that makes no contribution to model learning. Deduplication is a common technique in modern backup systems to reduce data redundancy. However, existing deduplication methods are invalid for training data. Hence, this paper proposes a novel deduplication strategy for the training data used for learning in a deep neural network classifier. Experimental results showed that the proposed deduplication strategy achieved 93% backup storage space reduction with only 1.3% loss of classification accuracy.

AB - The advent of the AI era has made it increasingly important to have an efficient backup system to protect training data from loss. Furthermore, a backup of the training data makes it possible to update or retrain the learned model as more data are collected. However, a huge backup overhead will result if a complete copy of all daily collected training data is always made to backup storage, especially because the data typically contain highly redundant information that makes no contribution to model learning. Deduplication is a common technique in modern backup systems to reduce data redundancy. However, existing deduplication methods are invalid for training data. Hence, this paper proposes a novel deduplication strategy for the training data used for learning in a deep neural network classifier. Experimental results showed that the proposed deduplication strategy achieved 93% backup storage space reduction with only 1.3% loss of classification accuracy.

KW - Backup systems

KW - Deduplication

KW - Deep learning

KW - Training data

UR - http://www.scopus.com/inward/record.url?scp=85098856011&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85098856011&partnerID=8YFLogxK

U2 - 10.1109/IPCCC50635.2020.9391523

DO - 10.1109/IPCCC50635.2020.9391523

M3 - Conference contribution

AN - SCOPUS:85098856011

T3 - 2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020

BT - 2020 IEEE 39th International Performance Computing and Communications Conference, IPCCC 2020

PB - Institute of Electrical and Electronics Engineers Inc.

Y2 - 6 November 2020 through 8 November 2020

ER -

AdaEmb-Encoder: Adaptive Embedding Spatial Encoder-Based Deduplication for Backing up Classifier Training Data

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access

OpenUrl availability

Other files and links

Fingerprint

Cite this