DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kim, Minkyung | ko |
dc.contributor.author | Kim, Junsik | ko |
dc.contributor.author | Yu, Jongmin | ko |
dc.contributor.author | Choi, Jun Kyun | ko |
dc.date.accessioned | 2023-09-20T06:00:18Z | - |
dc.date.available | 2023-09-20T06:00:18Z | - |
dc.date.created | 2023-09-20 | - |
dc.date.issued | 2022-11 | - |
dc.identifier.citation | 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022, pp.39 - 46 | - |
dc.identifier.issn | 2375-9232 | - |
dc.identifier.uri | http://hdl.handle.net/10203/312785 | - |
dc.description.abstract | One-class classification has been a prevailing method in building deep anomaly detection models under the assumption that a dataset consisting of normal samples is available. In practice, however, abnormal samples are often mixed in a training dataset, and they detrimentally affect the training of deep models, which limits their applicability. For robust normality learning of deep practical models, we propose an unsupervised deep one-class classification that learns normality from pseudo-labeled normal samples, i.e., outlier detection in single cluster scenarios. To this end, we propose a pseudo-labeling method by an adaptive threshold selected by ranking-based training dynamics. The experiments on 10 anomaly detection benchmarks show that our method effectively improves performance on anomaly detection by sizable margins. | - |
dc.language | English | - |
dc.publisher | IEEE Computer Society | - |
dc.title | Unsupervised Deep One-Class Classification with Adaptive Threshold based on Training Dynamics | - |
dc.type | Conference | - |
dc.identifier.wosid | 000971492200006 | - |
dc.identifier.scopusid | 2-s2.0-85148449205 | - |
dc.type.rims | CONF | - |
dc.citation.beginningpage | 39 | - |
dc.citation.endingpage | 46 | - |
dc.citation.publicationname | 22nd IEEE International Conference on Data Mining Workshops, ICDMW 2022 | - |
dc.identifier.conferencecountry | US | - |
dc.identifier.conferencelocation | Orlando, FL | - |
dc.identifier.doi | 10.1109/ICDMW58026.2022.00014 | - |
dc.contributor.localauthor | Choi, Jun Kyun | - |
dc.contributor.nonIdAuthor | Kim, Junsik | - |
dc.contributor.nonIdAuthor | Yu, Jongmin | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.