Matches in SemOpenAlex for { <https://semopenalex.org/work/W3113123028> ?p ?o ?g. }
Showing items 1 to 92 of
92
with 100 items per page.
- W3113123028 abstract "Self-supervised learning (SSL) is an efficient approach that addresses the issue of annotation shortage. The key part in SSL is its proxy task that defines the supervisory signals and drives the learning toward effective feature representations. However, most SSL approaches usually focus on a single proxy task, which greatly limits the expressive power of the learned features and therefore deteriorates the network generalization capacity. In this regard, we hereby propose three strategies of aggregation in terms of complementarity of various forms to boost the robustness of self-supervised learned features. In spatial context aggregative SSL, we contribute a heuristic SSL method that integrates two ad-hoc proxy tasks with spatial context complementarity, modeling global and local contextual features, respectively. We then propose a principled framework of multi-task aggregative self-supervised learning to form a unified representation, with an intent of exploiting feature complementarity among different tasks. Finally, in self-aggregative SSL, we propose to self-complement an existing proxy task with an auxiliary loss function based on a linear centered kernel alignment metric, which explicitly promotes the exploring of where are uncovered by the features learned from a proxy task at hand to further boost the modeling capability. Our extensive experiments on 2D natural image and 3D medical image classification tasks under limited annotation scenarios confirm that the proposed aggregation strategies successfully boost the classification accuracy." @default.
- W3113123028 created "2020-12-21" @default.
- W3113123028 creator A5018349566 @default.
- W3113123028 creator A5045656118 @default.
- W3113123028 creator A5051870558 @default.
- W3113123028 date "2020-12-14" @default.
- W3113123028 modified "2023-10-01" @default.
- W3113123028 title "Aggregative Self-Supervised Feature Learning." @default.
- W3113123028 cites W2321533354 @default.
- W3113123028 cites W2599837529 @default.
- W3113123028 cites W2913939497 @default.
- W3113123028 cites W2949517790 @default.
- W3113123028 cites W2949558627 @default.
- W3113123028 cites W2962742544 @default.
- W3113123028 cites W2962824366 @default.
- W3113123028 cites W2963420272 @default.
- W3113123028 cites W2963759780 @default.
- W3113123028 cites W2964060161 @default.
- W3113123028 cites W2970241862 @default.
- W3113123028 cites W2979888373 @default.
- W3113123028 cites W2979954119 @default.
- W3113123028 cites W3005680577 @default.
- W3113123028 cites W3033671339 @default.
- W3113123028 cites W3034559532 @default.
- W3113123028 cites W3034915791 @default.
- W3113123028 cites W3036982689 @default.
- W3113123028 cites W71154278 @default.
- W3113123028 hasPublicationYear "2020" @default.
- W3113123028 type Work @default.
- W3113123028 sameAs 3113123028 @default.
- W3113123028 citedByCount "0" @default.
- W3113123028 crossrefType "posted-content" @default.
- W3113123028 hasAuthorship W3113123028A5018349566 @default.
- W3113123028 hasAuthorship W3113123028A5045656118 @default.
- W3113123028 hasAuthorship W3113123028A5051870558 @default.
- W3113123028 hasConcept C104317684 @default.
- W3113123028 hasConcept C119857082 @default.
- W3113123028 hasConcept C153180895 @default.
- W3113123028 hasConcept C154945302 @default.
- W3113123028 hasConcept C185592680 @default.
- W3113123028 hasConcept C202269582 @default.
- W3113123028 hasConcept C22019652 @default.
- W3113123028 hasConcept C41008148 @default.
- W3113123028 hasConcept C50644808 @default.
- W3113123028 hasConcept C54355233 @default.
- W3113123028 hasConcept C55493867 @default.
- W3113123028 hasConcept C59404180 @default.
- W3113123028 hasConcept C63479239 @default.
- W3113123028 hasConcept C83665646 @default.
- W3113123028 hasConcept C86803240 @default.
- W3113123028 hasConceptScore W3113123028C104317684 @default.
- W3113123028 hasConceptScore W3113123028C119857082 @default.
- W3113123028 hasConceptScore W3113123028C153180895 @default.
- W3113123028 hasConceptScore W3113123028C154945302 @default.
- W3113123028 hasConceptScore W3113123028C185592680 @default.
- W3113123028 hasConceptScore W3113123028C202269582 @default.
- W3113123028 hasConceptScore W3113123028C22019652 @default.
- W3113123028 hasConceptScore W3113123028C41008148 @default.
- W3113123028 hasConceptScore W3113123028C50644808 @default.
- W3113123028 hasConceptScore W3113123028C54355233 @default.
- W3113123028 hasConceptScore W3113123028C55493867 @default.
- W3113123028 hasConceptScore W3113123028C59404180 @default.
- W3113123028 hasConceptScore W3113123028C63479239 @default.
- W3113123028 hasConceptScore W3113123028C83665646 @default.
- W3113123028 hasConceptScore W3113123028C86803240 @default.
- W3113123028 hasLocation W31131230281 @default.
- W3113123028 hasOpenAccess W3113123028 @default.
- W3113123028 hasPrimaryLocation W31131230281 @default.
- W3113123028 hasRelatedWork W2793856863 @default.
- W3113123028 hasRelatedWork W2950113471 @default.
- W3113123028 hasRelatedWork W2963465221 @default.
- W3113123028 hasRelatedWork W2977981067 @default.
- W3113123028 hasRelatedWork W2982754279 @default.
- W3113123028 hasRelatedWork W3021757253 @default.
- W3113123028 hasRelatedWork W3035468615 @default.
- W3113123028 hasRelatedWork W3044638916 @default.
- W3113123028 hasRelatedWork W3086550652 @default.
- W3113123028 hasRelatedWork W3090425949 @default.
- W3113123028 hasRelatedWork W3092805083 @default.
- W3113123028 hasRelatedWork W3135460631 @default.
- W3113123028 hasRelatedWork W3136903696 @default.
- W3113123028 hasRelatedWork W3145682503 @default.
- W3113123028 hasRelatedWork W3156968278 @default.
- W3113123028 hasRelatedWork W3160514115 @default.
- W3113123028 hasRelatedWork W3183977392 @default.
- W3113123028 hasRelatedWork W3185459701 @default.
- W3113123028 hasRelatedWork W3196984783 @default.
- W3113123028 hasRelatedWork W3202747382 @default.
- W3113123028 isParatext "false" @default.
- W3113123028 isRetracted "false" @default.
- W3113123028 magId "3113123028" @default.
- W3113123028 workType "article" @default.