Matches in SemOpenAlex for { <https://semopenalex.org/work/W4309590903> ?p ?o ?g. }
Showing items 1 to 73 of
73
with 100 items per page.
- W4309590903 abstract "The widespread success of convolutional neural networks may largely be attributed to their intrinsic property of translation equivariance. However, convolutions are not equivariant to variations in scale and fail to generalize to objects of different sizes. Despite recent advances in this field, it remains unclear how well current methods generalize to unobserved scales on real-world data and to what extent scale equivariance plays a role. To address this, we propose the novel Scaled and Translated Image Recognition (STIR) benchmark based on four different domains. Additionally, we introduce a new family of models that applies many re-scaled kernels with shared weights in parallel and then selects the most appropriate one. Our experimental results on STIR show that both the existing and proposed approaches can improve generalization across scales compared to standard convolutions. We also demonstrate that our family of models is able to generalize well towards larger scales and improve scale equivariance. Moreover, due to their unique design we can validate that kernel selection is consistent with input scale. Even so, none of the evaluated models maintain their performance for large differences in scale, demonstrating that a general understanding of how scale equivariance can improve generalization and robustness is still lacking." @default.
- W4309590903 created "2022-11-28" @default.
- W4309590903 creator A5009875578 @default.
- W4309590903 creator A5014144494 @default.
- W4309590903 creator A5023277657 @default.
- W4309590903 creator A5023645490 @default.
- W4309590903 creator A5028233502 @default.
- W4309590903 creator A5066328216 @default.
- W4309590903 creator A5078457885 @default.
- W4309590903 date "2022-11-18" @default.
- W4309590903 modified "2023-10-14" @default.
- W4309590903 title "Just a Matter of Scale? Reevaluating Scale Equivariance in Convolutional Neural Networks" @default.
- W4309590903 doi "https://doi.org/10.48550/arxiv.2211.10288" @default.
- W4309590903 hasPublicationYear "2022" @default.
- W4309590903 type Work @default.
- W4309590903 citedByCount "0" @default.
- W4309590903 crossrefType "posted-content" @default.
- W4309590903 hasAuthorship W4309590903A5009875578 @default.
- W4309590903 hasAuthorship W4309590903A5014144494 @default.
- W4309590903 hasAuthorship W4309590903A5023277657 @default.
- W4309590903 hasAuthorship W4309590903A5023645490 @default.
- W4309590903 hasAuthorship W4309590903A5028233502 @default.
- W4309590903 hasAuthorship W4309590903A5066328216 @default.
- W4309590903 hasAuthorship W4309590903A5078457885 @default.
- W4309590903 hasBestOaLocation W43095909031 @default.
- W4309590903 hasConcept C104317684 @default.
- W4309590903 hasConcept C134306372 @default.
- W4309590903 hasConcept C154945302 @default.
- W4309590903 hasConcept C171036898 @default.
- W4309590903 hasConcept C177148314 @default.
- W4309590903 hasConcept C185592680 @default.
- W4309590903 hasConcept C202444582 @default.
- W4309590903 hasConcept C205649164 @default.
- W4309590903 hasConcept C2778755073 @default.
- W4309590903 hasConcept C33923547 @default.
- W4309590903 hasConcept C41008148 @default.
- W4309590903 hasConcept C55493867 @default.
- W4309590903 hasConcept C58640448 @default.
- W4309590903 hasConcept C63479239 @default.
- W4309590903 hasConcept C74193536 @default.
- W4309590903 hasConcept C81363708 @default.
- W4309590903 hasConceptScore W4309590903C104317684 @default.
- W4309590903 hasConceptScore W4309590903C134306372 @default.
- W4309590903 hasConceptScore W4309590903C154945302 @default.
- W4309590903 hasConceptScore W4309590903C171036898 @default.
- W4309590903 hasConceptScore W4309590903C177148314 @default.
- W4309590903 hasConceptScore W4309590903C185592680 @default.
- W4309590903 hasConceptScore W4309590903C202444582 @default.
- W4309590903 hasConceptScore W4309590903C205649164 @default.
- W4309590903 hasConceptScore W4309590903C2778755073 @default.
- W4309590903 hasConceptScore W4309590903C33923547 @default.
- W4309590903 hasConceptScore W4309590903C41008148 @default.
- W4309590903 hasConceptScore W4309590903C55493867 @default.
- W4309590903 hasConceptScore W4309590903C58640448 @default.
- W4309590903 hasConceptScore W4309590903C63479239 @default.
- W4309590903 hasConceptScore W4309590903C74193536 @default.
- W4309590903 hasConceptScore W4309590903C81363708 @default.
- W4309590903 hasLocation W43095909031 @default.
- W4309590903 hasOpenAccess W4309590903 @default.
- W4309590903 hasPrimaryLocation W43095909031 @default.
- W4309590903 hasRelatedWork W2735477435 @default.
- W4309590903 hasRelatedWork W2748454020 @default.
- W4309590903 hasRelatedWork W3016958897 @default.
- W4309590903 hasRelatedWork W3045739591 @default.
- W4309590903 hasRelatedWork W3119610945 @default.
- W4309590903 hasRelatedWork W3181746755 @default.
- W4309590903 hasRelatedWork W4283315865 @default.
- W4309590903 hasRelatedWork W4283379348 @default.
- W4309590903 hasRelatedWork W4312417841 @default.
- W4309590903 hasRelatedWork W4385415357 @default.
- W4309590903 isParatext "false" @default.
- W4309590903 isRetracted "false" @default.
- W4309590903 workType "article" @default.