Publications
Detailed Information
Continual Learning on Noisy Data Streams via Self-Purified Replay
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Kim, Chris Dongjoo | - |
dc.contributor.author | Jeong, Jinseo | - |
dc.contributor.author | Moon, Sangwoo | - |
dc.contributor.author | Kim, Gun Hee | - |
dc.date.accessioned | 2022-06-24T00:26:25Z | - |
dc.date.available | 2022-06-24T00:26:25Z | - |
dc.date.created | 2022-05-09 | - |
dc.date.issued | 2021-01 | - |
dc.identifier.citation | Proceedings of the IEEE International Conference on Computer Vision, pp.517-527 | - |
dc.identifier.issn | 1550-5499 | - |
dc.identifier.uri | https://hdl.handle.net/10371/183772 | - |
dc.description.abstract | © 2021 IEEEContinually learning in the real world must overcome many challenges, among which noisy labels are a common and inevitable issue. In this work, we present a replay-based continual learning framework that simultaneously addresses both catastrophic forgetting and noisy labels for the first time. Our solution is based on two observations; (i) forgetting can be mitigated even with noisy labels via self-supervised learning, and (ii) the purity of the replay buffer is crucial. Building on this regard, we propose two key components of our method: (i) a self-supervised replay technique named Self-Replay which can circumvent erroneous training signals arising from noisy labeled data, and (ii) the Self-Centered filter that maintains a purified replay buffer via centrality-based stochastic graph ensembles. The empirical results on MNIST, CIFAR-10, CIFAR-100, and WebVision with real-world noise demonstrate that our framework can maintain a highly pure replay buffer amidst noisy streamed data while greatly outperforming the combinations of the state-of-the-art continual learning and noisy label learning methods. | - |
dc.language | 영어 | - |
dc.publisher | IEEE | - |
dc.title | Continual Learning on Noisy Data Streams via Self-Purified Replay | - |
dc.type | Article | - |
dc.identifier.doi | 10.1109/ICCV48922.2021.00058 | - |
dc.citation.journaltitle | Proceedings of the IEEE International Conference on Computer Vision | - |
dc.identifier.scopusid | 2-s2.0-85127025243 | - |
dc.citation.endpage | 527 | - |
dc.citation.startpage | 517 | - |
dc.description.isOpenAccess | N | - |
dc.contributor.affiliatedAuthor | Kim, Gun Hee | - |
dc.type.docType | Conference Paper | - |
dc.description.journalClass | 1 | - |
- Appears in Collections:
- Files in This Item:
- There are no files associated with this item.
Item View & Download Count
Items in S-Space are protected by copyright, with all rights reserved, unless otherwise indicated.