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
현재까지 많은 연구에서 레이블이 지정되지 않은 데이터를 분류하기 위해 클러스터링을 사용했습니다. Deep Separate Clustering은 기존 클러스터링 알고리즘에 여러 딥러닝 모델을 적용하여 클러스터의 분포를 보다 명확하게 분리합니다. 본 논문에서는 입력 이미지의 특징을 학습하기 위해 Convolutional Autoencoder를 사용합니다. 이에 따라, k- 컨벌루션 오토인코더가 학습한 인코딩된 계층 기능을 사용하여 클러스터링이 수행됨을 의미합니다. 그런 다음 중앙 손실 함수를 추가하여 데이터 포인트를 클러스터로 집계하여 클러스터 내 동질성을 높입니다. 마지막으로 클러스터 간 분리성을 계산하고 증가시킵니다. 모든 손실 함수를 단일 전역 목적 함수로 결합합니다. 우리의 새로운 심층 클러스터링 방법은 동일한 조건에서의 실험에서 비교할 때 기존 클러스터링 접근 방식의 성능을 능가합니다.
Byeonghak KIM
Korea University
Murray LOEW
George Washington University
David K. HAN
Drexel University
Hanseok KO
Korea University
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.
부
Byeonghak KIM, Murray LOEW, David K. HAN, Hanseok KO, "Deep Clustering for Improved Inter-Cluster Separability and Intra-Cluster Homogeneity with Cohesive Loss" in IEICE TRANSACTIONS on Information,
vol. E104-D, no. 5, pp. 776-780, May 2021, doi: 10.1587/transinf.2020EDL8138.
Abstract: To date, many studies have employed clustering for the classification of unlabeled data. Deep separate clustering applies several deep learning models to conventional clustering algorithms to more clearly separate the distribution of the clusters. In this paper, we employ a convolutional autoencoder to learn the features of input images. Following this, k-means clustering is conducted using the encoded layer features learned by the convolutional autoencoder. A center loss function is then added to aggregate the data points into clusters to increase the intra-cluster homogeneity. Finally, we calculate and increase the inter-cluster separability. We combine all loss functions into a single global objective function. Our new deep clustering method surpasses the performance of existing clustering approaches when compared in experiments under the same conditions.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2020EDL8138/_p
부
@ARTICLE{e104-d_5_776,
author={Byeonghak KIM, Murray LOEW, David K. HAN, Hanseok KO, },
journal={IEICE TRANSACTIONS on Information},
title={Deep Clustering for Improved Inter-Cluster Separability and Intra-Cluster Homogeneity with Cohesive Loss},
year={2021},
volume={E104-D},
number={5},
pages={776-780},
abstract={To date, many studies have employed clustering for the classification of unlabeled data. Deep separate clustering applies several deep learning models to conventional clustering algorithms to more clearly separate the distribution of the clusters. In this paper, we employ a convolutional autoencoder to learn the features of input images. Following this, k-means clustering is conducted using the encoded layer features learned by the convolutional autoencoder. A center loss function is then added to aggregate the data points into clusters to increase the intra-cluster homogeneity. Finally, we calculate and increase the inter-cluster separability. We combine all loss functions into a single global objective function. Our new deep clustering method surpasses the performance of existing clustering approaches when compared in experiments under the same conditions.},
keywords={},
doi={10.1587/transinf.2020EDL8138},
ISSN={1745-1361},
month={May},}
부
TY - JOUR
TI - Deep Clustering for Improved Inter-Cluster Separability and Intra-Cluster Homogeneity with Cohesive Loss
T2 - IEICE TRANSACTIONS on Information
SP - 776
EP - 780
AU - Byeonghak KIM
AU - Murray LOEW
AU - David K. HAN
AU - Hanseok KO
PY - 2021
DO - 10.1587/transinf.2020EDL8138
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E104-D
IS - 5
JA - IEICE TRANSACTIONS on Information
Y1 - May 2021
AB - To date, many studies have employed clustering for the classification of unlabeled data. Deep separate clustering applies several deep learning models to conventional clustering algorithms to more clearly separate the distribution of the clusters. In this paper, we employ a convolutional autoencoder to learn the features of input images. Following this, k-means clustering is conducted using the encoded layer features learned by the convolutional autoencoder. A center loss function is then added to aggregate the data points into clusters to increase the intra-cluster homogeneity. Finally, we calculate and increase the inter-cluster separability. We combine all loss functions into a single global objective function. Our new deep clustering method surpasses the performance of existing clustering approaches when compared in experiments under the same conditions.
ER -