Matches in SemOpenAlex for { <https://semopenalex.org/work/W4382240131> ?p ?o ?g. }
Showing items 1 to 67 of
67
with 100 items per page.
- W4382240131 endingPage "3524" @default.
- W4382240131 startingPage "3516" @default.
- W4382240131 abstract "Positional encoding is important for vision transformer (ViT) to capture the spatial structure of the input image. General effectiveness has been proven in ViT. In our work we propose to train ViT to recognize the positional label of patches of the input image, this apparently simple task actually yields a meaningful self-supervisory task. Based on previous work on ViT positional encoding, we propose two positional labels dedicated to 2D images including absolute position and relative position. Our positional labels can be easily plugged into various current ViT variants. It can work in two ways: (a) As an auxiliary training target for vanilla ViT for better performance. (b) Combine the self-supervised ViT to provide a more powerful self-supervised signal for semantic feature learning. Experiments demonstrate that with the proposed self-supervised methods, ViT-B and Swin-B gain improvements of 1.20% (top-1 Acc) and 0.74% (top-1 Acc) on ImageNet, respectively, and 6.15% and 1.14% improvement on Mini-ImageNet. The code is publicly available at: https://github.com/zhangzhemin/PositionalLabel." @default.
- W4382240131 created "2023-06-28" @default.
- W4382240131 creator A5000673554 @default.
- W4382240131 creator A5009729539 @default.
- W4382240131 date "2023-06-26" @default.
- W4382240131 modified "2023-09-23" @default.
- W4382240131 title "Positional Label for Self-Supervised Vision Transformer" @default.
- W4382240131 doi "https://doi.org/10.1609/aaai.v37i3.25461" @default.
- W4382240131 hasPublicationYear "2023" @default.
- W4382240131 type Work @default.
- W4382240131 citedByCount "0" @default.
- W4382240131 crossrefType "journal-article" @default.
- W4382240131 hasAuthorship W4382240131A5000673554 @default.
- W4382240131 hasAuthorship W4382240131A5009729539 @default.
- W4382240131 hasBestOaLocation W43822401311 @default.
- W4382240131 hasConcept C119599485 @default.
- W4382240131 hasConcept C125411270 @default.
- W4382240131 hasConcept C127413603 @default.
- W4382240131 hasConcept C136389625 @default.
- W4382240131 hasConcept C138885662 @default.
- W4382240131 hasConcept C153180895 @default.
- W4382240131 hasConcept C154945302 @default.
- W4382240131 hasConcept C165801399 @default.
- W4382240131 hasConcept C201995342 @default.
- W4382240131 hasConcept C2776401178 @default.
- W4382240131 hasConcept C2780451532 @default.
- W4382240131 hasConcept C31972630 @default.
- W4382240131 hasConcept C41008148 @default.
- W4382240131 hasConcept C41895202 @default.
- W4382240131 hasConcept C50644808 @default.
- W4382240131 hasConcept C66322947 @default.
- W4382240131 hasConceptScore W4382240131C119599485 @default.
- W4382240131 hasConceptScore W4382240131C125411270 @default.
- W4382240131 hasConceptScore W4382240131C127413603 @default.
- W4382240131 hasConceptScore W4382240131C136389625 @default.
- W4382240131 hasConceptScore W4382240131C138885662 @default.
- W4382240131 hasConceptScore W4382240131C153180895 @default.
- W4382240131 hasConceptScore W4382240131C154945302 @default.
- W4382240131 hasConceptScore W4382240131C165801399 @default.
- W4382240131 hasConceptScore W4382240131C201995342 @default.
- W4382240131 hasConceptScore W4382240131C2776401178 @default.
- W4382240131 hasConceptScore W4382240131C2780451532 @default.
- W4382240131 hasConceptScore W4382240131C31972630 @default.
- W4382240131 hasConceptScore W4382240131C41008148 @default.
- W4382240131 hasConceptScore W4382240131C41895202 @default.
- W4382240131 hasConceptScore W4382240131C50644808 @default.
- W4382240131 hasConceptScore W4382240131C66322947 @default.
- W4382240131 hasIssue "3" @default.
- W4382240131 hasLocation W43822401311 @default.
- W4382240131 hasOpenAccess W4382240131 @default.
- W4382240131 hasPrimaryLocation W43822401311 @default.
- W4382240131 hasRelatedWork W1504288058 @default.
- W4382240131 hasRelatedWork W2017205855 @default.
- W4382240131 hasRelatedWork W2048505601 @default.
- W4382240131 hasRelatedWork W2116675934 @default.
- W4382240131 hasRelatedWork W2167293474 @default.
- W4382240131 hasRelatedWork W2331674254 @default.
- W4382240131 hasRelatedWork W2544359817 @default.
- W4382240131 hasRelatedWork W2979079341 @default.
- W4382240131 hasRelatedWork W3042897387 @default.
- W4382240131 hasRelatedWork W4310007291 @default.
- W4382240131 hasVolume "37" @default.
- W4382240131 isParatext "false" @default.
- W4382240131 isRetracted "false" @default.
- W4382240131 workType "article" @default.