A Variable Memory Length Auto Encoder

Mohomad Mohomad Ibunu; Samuel Weller; Clive Cheong Took

doi:10.1109/IJCNN52387.2021.9533475

A Variable Memory Length Auto Encoder

Mohomad Mohomad Ibunu, Samuel Weller, Clive Cheong Took

Research output: Contribution to conference › Paper › peer-review

129 Downloads (Pure)

Abstract

Auto-encoders typically require batch learning to be effective. There is a lack of online learning mechanisms for auto-encoders. To address this shortcoming in the literature, we propose an auto-encoder that can not only learn on a sample-by-sample basis without back-propagation but also has a memory to benefit from past learning. The memory can be adapted to fit the current state of the data by varying the memory length of the auto-encoder. Simulation supports our approach, especially when the data is nonstationary.

Original language	English
DOIs	https://doi.org/10.1109/IJCNN52387.2021.9533475
Publication status	Published - Jul 2021
Event	The International Joint Conference on Neural Networks 2021 - Virtual Duration: 18 Jul 2021 → 22 Jul 2021 https://www.ijcnn.org/

Conference

Conference	The International Joint Conference on Neural Networks 2021
Abbreviated title	IJCNN 2021
Period	18/07/21 → 22/07/21
Internet address	https://www.ijcnn.org/

Access to Document

10.1109/IJCNN52387.2021.9533475

Accepted ManuscriptAccepted author manuscript, 1.03 MB

Cite this

@conference{12f10b30d45f4900836cca334fbd59d4,

title = "A Variable Memory Length Auto Encoder",

abstract = "Auto-encoders typically require batch learning to be effective. There is a lack of online learning mechanisms for auto-encoders. To address this shortcoming in the literature, we propose an auto-encoder that can not only learn on a sample-by-sample basis without back-propagation but also has a memory to benefit from past learning. The memory can be adapted to fit the current state of the data by varying the memory length of the auto-encoder. Simulation supports our approach, especially when the data is nonstationary.",

author = "{Mohomad Ibunu}, Mohomad and Samuel Weller and {Cheong Took}, Clive",

year = "2021",

month = jul,

doi = "10.1109/IJCNN52387.2021.9533475",

language = "English",

note = "The International Joint Conference on Neural Networks 2021, IJCNN 2021 ; Conference date: 18-07-2021 Through 22-07-2021",

url = "https://www.ijcnn.org/",

}

TY - CONF

T1 - A Variable Memory Length Auto Encoder

AU - Mohomad Ibunu, Mohomad

AU - Weller, Samuel

AU - Cheong Took, Clive

PY - 2021/7

Y1 - 2021/7

N2 - Auto-encoders typically require batch learning to be effective. There is a lack of online learning mechanisms for auto-encoders. To address this shortcoming in the literature, we propose an auto-encoder that can not only learn on a sample-by-sample basis without back-propagation but also has a memory to benefit from past learning. The memory can be adapted to fit the current state of the data by varying the memory length of the auto-encoder. Simulation supports our approach, especially when the data is nonstationary.

AB - Auto-encoders typically require batch learning to be effective. There is a lack of online learning mechanisms for auto-encoders. To address this shortcoming in the literature, we propose an auto-encoder that can not only learn on a sample-by-sample basis without back-propagation but also has a memory to benefit from past learning. The memory can be adapted to fit the current state of the data by varying the memory length of the auto-encoder. Simulation supports our approach, especially when the data is nonstationary.

UR - https://www.ijcnn.org/draft-program-ijcnn-2021

U2 - 10.1109/IJCNN52387.2021.9533475

DO - 10.1109/IJCNN52387.2021.9533475

M3 - Paper

T2 - The International Joint Conference on Neural Networks 2021

Y2 - 18 July 2021 through 22 July 2021

ER -

A Variable Memory Length Auto Encoder

Abstract

Conference

Access to Document

Other files and links

Cite this