The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
O Aprendizado Incremental, uma metodologia de aprendizado de máquina, treina os dados de entrada que chegam continuamente e amplia o conhecimento do modelo. Quando se trata de fluxos de dados não rotulados, a tarefa de aprendizagem incremental torna-se mais desafiadora. Nossa metodologia de aprendizagem incremental recentemente proposta, Data Augmented Incremental Learning (DIÁRIO), aprende os fluxos em tempo real cada vez maiores com recursos e tempo de memória reduzidos. Inicialmente, os lotes não rotulados de fluxos de dados são agrupados usando o algoritmo de agrupamento proposto, Clustering baseado em Autoencoder e Modelo Gaussiano (CLAG). Mais tarde, DIÁRIO cria um modelo incremental atualizado para os clusters rotulados usando aumento de dados. DIÁRIO evita o retreinamento de amostras antigas e retém apenas o modelo incremental atualizado mais recentemente, contendo todas as informações da classe antiga. O uso de aumento de dados em DIÁRIO combina os clusters semelhantes gerados com diferentes lotes de dados. Uma série de experimentos comprovou o desempenho significativo de CLAG e DIÁRIO, produzindo um modelo incremental escalável e eficiente.
Sathya MADHUSUDHANAN
Sri Sivasubramaniya Nadar College of Engineering
Suresh JAGANATHAN
Sri Sivasubramaniya Nadar College of Engineering
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copiar
Sathya MADHUSUDHANAN, Suresh JAGANATHAN, "Data Augmented Incremental Learning (DAIL) for Unsupervised Data" in IEICE TRANSACTIONS on Information,
vol. E105-D, no. 6, pp. 1185-1195, June 2022, doi: 10.1587/transinf.2021EDP7213.
Abstract: Incremental Learning, a machine learning methodology, trains the continuously arriving input data and extends the model's knowledge. When it comes to unlabeled data streams, incremental learning task becomes more challenging. Our newly proposed incremental learning methodology, Data Augmented Incremental Learning (DAIL), learns the ever-increasing real-time streams with reduced memory resources and time. Initially, the unlabeled batches of data streams are clustered using the proposed clustering algorithm, Clustering based on Autoencoder and Gaussian Model (CLAG). Later, DAIL creates an updated incremental model for the labelled clusters using data augmentation. DAIL avoids the retraining of old samples and retains only the most recently updated incremental model holding all old class information. The use of data augmentation in DAIL combines the similar clusters generated with different data batches. A series of experiments verified the significant performance of CLAG and DAIL, producing scalable and efficient incremental model.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2021EDP7213/_p
Copiar
@ARTICLE{e105-d_6_1185,
author={Sathya MADHUSUDHANAN, Suresh JAGANATHAN, },
journal={IEICE TRANSACTIONS on Information},
title={Data Augmented Incremental Learning (DAIL) for Unsupervised Data},
year={2022},
volume={E105-D},
number={6},
pages={1185-1195},
abstract={Incremental Learning, a machine learning methodology, trains the continuously arriving input data and extends the model's knowledge. When it comes to unlabeled data streams, incremental learning task becomes more challenging. Our newly proposed incremental learning methodology, Data Augmented Incremental Learning (DAIL), learns the ever-increasing real-time streams with reduced memory resources and time. Initially, the unlabeled batches of data streams are clustered using the proposed clustering algorithm, Clustering based on Autoencoder and Gaussian Model (CLAG). Later, DAIL creates an updated incremental model for the labelled clusters using data augmentation. DAIL avoids the retraining of old samples and retains only the most recently updated incremental model holding all old class information. The use of data augmentation in DAIL combines the similar clusters generated with different data batches. A series of experiments verified the significant performance of CLAG and DAIL, producing scalable and efficient incremental model.},
keywords={},
doi={10.1587/transinf.2021EDP7213},
ISSN={1745-1361},
month={June},}
Copiar
TY - JOUR
TI - Data Augmented Incremental Learning (DAIL) for Unsupervised Data
T2 - IEICE TRANSACTIONS on Information
SP - 1185
EP - 1195
AU - Sathya MADHUSUDHANAN
AU - Suresh JAGANATHAN
PY - 2022
DO - 10.1587/transinf.2021EDP7213
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E105-D
IS - 6
JA - IEICE TRANSACTIONS on Information
Y1 - June 2022
AB - Incremental Learning, a machine learning methodology, trains the continuously arriving input data and extends the model's knowledge. When it comes to unlabeled data streams, incremental learning task becomes more challenging. Our newly proposed incremental learning methodology, Data Augmented Incremental Learning (DAIL), learns the ever-increasing real-time streams with reduced memory resources and time. Initially, the unlabeled batches of data streams are clustered using the proposed clustering algorithm, Clustering based on Autoencoder and Gaussian Model (CLAG). Later, DAIL creates an updated incremental model for the labelled clusters using data augmentation. DAIL avoids the retraining of old samples and retains only the most recently updated incremental model holding all old class information. The use of data augmentation in DAIL combines the similar clusters generated with different data batches. A series of experiments verified the significant performance of CLAG and DAIL, producing scalable and efficient incremental model.
ER -