Dataset condensation with atypical samples특이한 샘플들을 이용한 데이타셋 압축

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 3
  • Download : 0
DC FieldValueLanguage
dc.contributor.advisor정혜원-
dc.contributor.authorLee, Yongmin-
dc.contributor.author이용민-
dc.date.accessioned2024-07-30T19:31:22Z-
dc.date.available2024-07-30T19:31:22Z-
dc.date.issued2024-
dc.identifier.urihttp://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=1096784&flag=dissertationen_US
dc.identifier.urihttp://hdl.handle.net/10203/321566-
dc.description학위논문(석사) - 한국과학기술원 : 전기및전자공학부, 2024.2,[iii, 17 p. :]-
dc.description.abstractThe advancement of deep learning relies on huge datasets, which involves extensive training cost. To reduce the size of dataset, dataset distillation emerges, which aims to synthesize a small dataset encoding the information of the entire training set. However, we argue that existing dataset condensation method excessively focus on synthesizing typical samples rather than atypical samples, thereby diminish the network’s generalization capacity. To address the limitation, we propose a new strategy, Select-and-Match (SelMatch), that combines dataset condensation and selection. Our method selects important samples first, and perform condensation with regard to selected samples, which enriches the overall information within the synthetic dataset. We evaluate our method on the CIFAR-100 benchmark and demonstrate its superiority over state-of-the-art selection-only and condensation-only methods.-
dc.languageeng-
dc.publisher한국과학기술원-
dc.subject데이터 효율적 학습▼a데이터셋 축소▼a최적화▼a데이터 선택-
dc.subjectData-efficient learning▼aDataset condensation▼aOptimization▼aSubset selection-
dc.titleDataset condensation with atypical samples-
dc.title.alternative특이한 샘플들을 이용한 데이타셋 압축-
dc.typeThesis(Master)-
dc.identifier.CNRN325007-
dc.description.department한국과학기술원 :전기및전자공학부,-
dc.contributor.alternativeauthorChung, Hye Won-
Appears in Collection
EE-Theses_Master(석사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0