Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386076260> ?p ?o ?g. }
- W4386076260 abstract "Deep learning in general domains has constantly been extended to domain-specific tasks requiring the recognition of fine-grained characteristics. However, real-world applications for fine-grained tasks suffer from two challenges: a high reliance on expert knowledge for annotation and necessity of a versatile model for various downstream tasks in a specific domain (e.g., prediction of categories, bounding boxes, or pixel-wise annotations). Fortunately, the recent self-supervised learning (SSL) is a promising approach to pretrain a model without annotations, serving as an effective initialization for any downstream tasks. Since SSL does not rely on the presence of annotation, in general, it utilizes the large-scale unlabeled dataset, referred to as an open-set. In this sense, we introduce a novel Open-Set Self-Supervised Learning problem under the assumption that a large-scale unlabeled open-set is available, as well as the fine-grained target dataset, during a pretraining phase. In our problem setup, it is crucial to consider the distribution mismatch between the open-set and target dataset. Hence, we propose SimCore algorithm to sample a coreset, the subset of an open-set that has a minimum distance to the target dataset in the latent space. We demonstrate that SimCore significantly improves representation learning performance through extensive experimental settings, including eleven fine-grained datasets and seven open-sets in various downstream tasks." @default.
- W4386076260 created "2023-08-23" @default.
- W4386076260 creator A5056038923 @default.
- W4386076260 creator A5061665831 @default.
- W4386076260 creator A5091674853 @default.
- W4386076260 date "2023-06-01" @default.
- W4386076260 modified "2023-09-27" @default.
- W4386076260 title "Coreset Sampling from Open-Set for Fine-Grained Self-Supervised Learning" @default.
- W4386076260 cites W1834627138 @default.
- W4386076260 cites W1977295328 @default.
- W4386076260 cites W2011674654 @default.
- W4386076260 cites W2038765747 @default.
- W4386076260 cites W2047643928 @default.
- W4386076260 cites W2108598243 @default.
- W4386076260 cites W2119880843 @default.
- W4386076260 cites W2124219775 @default.
- W4386076260 cites W2138011018 @default.
- W4386076260 cites W2152161678 @default.
- W4386076260 cites W2194775991 @default.
- W4386076260 cites W2533598788 @default.
- W4386076260 cites W2549139847 @default.
- W4386076260 cites W2797977484 @default.
- W4386076260 cites W2963149653 @default.
- W4386076260 cites W2963351448 @default.
- W4386076260 cites W2965989958 @default.
- W4386076260 cites W2986821660 @default.
- W4386076260 cites W2997131443 @default.
- W4386076260 cites W3034521057 @default.
- W4386076260 cites W3035524453 @default.
- W4386076260 cites W3102616566 @default.
- W4386076260 cites W3114632476 @default.
- W4386076260 cites W3134961575 @default.
- W4386076260 cites W3159481202 @default.
- W4386076260 cites W3160314846 @default.
- W4386076260 cites W3161438416 @default.
- W4386076260 cites W3171007011 @default.
- W4386076260 cites W3173329034 @default.
- W4386076260 cites W3175593095 @default.
- W4386076260 cites W3176276772 @default.
- W4386076260 cites W3193137605 @default.
- W4386076260 cites W3216968833 @default.
- W4386076260 cites W4312595781 @default.
- W4386076260 cites W4313156423 @default.
- W4386076260 doi "https://doi.org/10.1109/cvpr52729.2023.00728" @default.
- W4386076260 hasPublicationYear "2023" @default.
- W4386076260 type Work @default.
- W4386076260 citedByCount "0" @default.
- W4386076260 crossrefType "proceedings-article" @default.
- W4386076260 hasAuthorship W4386076260A5056038923 @default.
- W4386076260 hasAuthorship W4386076260A5061665831 @default.
- W4386076260 hasAuthorship W4386076260A5091674853 @default.
- W4386076260 hasConcept C114466953 @default.
- W4386076260 hasConcept C118615104 @default.
- W4386076260 hasConcept C119857082 @default.
- W4386076260 hasConcept C121332964 @default.
- W4386076260 hasConcept C124101348 @default.
- W4386076260 hasConcept C136389625 @default.
- W4386076260 hasConcept C154945302 @default.
- W4386076260 hasConcept C177264268 @default.
- W4386076260 hasConcept C17744445 @default.
- W4386076260 hasConcept C199360897 @default.
- W4386076260 hasConcept C199539241 @default.
- W4386076260 hasConcept C2776321320 @default.
- W4386076260 hasConcept C2776359362 @default.
- W4386076260 hasConcept C2778755073 @default.
- W4386076260 hasConcept C33923547 @default.
- W4386076260 hasConcept C41008148 @default.
- W4386076260 hasConcept C42357961 @default.
- W4386076260 hasConcept C50644808 @default.
- W4386076260 hasConcept C59404180 @default.
- W4386076260 hasConcept C62520636 @default.
- W4386076260 hasConcept C63584917 @default.
- W4386076260 hasConcept C94625758 @default.
- W4386076260 hasConceptScore W4386076260C114466953 @default.
- W4386076260 hasConceptScore W4386076260C118615104 @default.
- W4386076260 hasConceptScore W4386076260C119857082 @default.
- W4386076260 hasConceptScore W4386076260C121332964 @default.
- W4386076260 hasConceptScore W4386076260C124101348 @default.
- W4386076260 hasConceptScore W4386076260C136389625 @default.
- W4386076260 hasConceptScore W4386076260C154945302 @default.
- W4386076260 hasConceptScore W4386076260C177264268 @default.
- W4386076260 hasConceptScore W4386076260C17744445 @default.
- W4386076260 hasConceptScore W4386076260C199360897 @default.
- W4386076260 hasConceptScore W4386076260C199539241 @default.
- W4386076260 hasConceptScore W4386076260C2776321320 @default.
- W4386076260 hasConceptScore W4386076260C2776359362 @default.
- W4386076260 hasConceptScore W4386076260C2778755073 @default.
- W4386076260 hasConceptScore W4386076260C33923547 @default.
- W4386076260 hasConceptScore W4386076260C41008148 @default.
- W4386076260 hasConceptScore W4386076260C42357961 @default.
- W4386076260 hasConceptScore W4386076260C50644808 @default.
- W4386076260 hasConceptScore W4386076260C59404180 @default.
- W4386076260 hasConceptScore W4386076260C62520636 @default.
- W4386076260 hasConceptScore W4386076260C63584917 @default.
- W4386076260 hasConceptScore W4386076260C94625758 @default.
- W4386076260 hasFunder F4320335489 @default.
- W4386076260 hasLocation W43860762601 @default.
- W4386076260 hasOpenAccess W4386076260 @default.
- W4386076260 hasPrimaryLocation W43860762601 @default.
- W4386076260 hasRelatedWork W2953328427 @default.