Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385225652> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W4385225652 abstract "Token interaction operation is one of the core modules in MLP-based models to exchange and aggregate information between different spatial locations. However, the power of token interaction on the spatial dimension is highly dependent on the spatial resolution of the feature maps, which limits the model's expressive ability, especially in deep layers where the feature are down-sampled to a small spatial size. To address this issue, we present a novel method called textbf{Strip-MLP} to enrich the token interaction power in three ways. Firstly, we introduce a new MLP paradigm called Strip MLP layer that allows the token to interact with other tokens in a cross-strip manner, enabling the tokens in a row (or column) to contribute to the information aggregations in adjacent but different strips of rows (or columns). Secondly, a textbf{C}ascade textbf{G}roup textbf{S}trip textbf{M}ixing textbf{M}odule (CGSMM) is proposed to overcome the performance degradation caused by small spatial feature size. The module allows tokens to interact more effectively in the manners of within-patch and cross-patch, which is independent to the feature spatial size. Finally, based on the Strip MLP layer, we propose a novel textbf{L}ocal textbf{S}trip textbf{M}ixing textbf{M}odule (LSMM) to boost the token interaction power in the local region. Extensive experiments demonstrate that Strip-MLP significantly improves the performance of MLP-based models on small datasets and obtains comparable or even better results on ImageNet. In particular, Strip-MLP models achieve higher average Top-1 accuracy than existing MLP-based models by +2.44% on Caltech-101 and +2.16% on CIFAR-100. The source codes will be available at~href{https://github.com/Med-Process/Strip_MLP{https://github.com/Med-Process/Strip_MLP}." @default.
- W4385225652 created "2023-07-25" @default.
- W4385225652 creator A5015744133 @default.
- W4385225652 creator A5030851967 @default.
- W4385225652 creator A5055543314 @default.
- W4385225652 creator A5060307752 @default.
- W4385225652 creator A5063818643 @default.
- W4385225652 creator A5071271092 @default.
- W4385225652 creator A5090914417 @default.
- W4385225652 date "2023-07-21" @default.
- W4385225652 modified "2023-09-23" @default.
- W4385225652 title "Strip-MLP: Efficient Token Interaction for Vision MLP" @default.
- W4385225652 doi "https://doi.org/10.48550/arxiv.2307.11458" @default.
- W4385225652 hasPublicationYear "2023" @default.
- W4385225652 type Work @default.
- W4385225652 citedByCount "0" @default.
- W4385225652 crossrefType "posted-content" @default.
- W4385225652 hasAuthorship W4385225652A5015744133 @default.
- W4385225652 hasAuthorship W4385225652A5030851967 @default.
- W4385225652 hasAuthorship W4385225652A5055543314 @default.
- W4385225652 hasAuthorship W4385225652A5060307752 @default.
- W4385225652 hasAuthorship W4385225652A5063818643 @default.
- W4385225652 hasAuthorship W4385225652A5071271092 @default.
- W4385225652 hasAuthorship W4385225652A5090914417 @default.
- W4385225652 hasBestOaLocation W43852256521 @default.
- W4385225652 hasConcept C11413529 @default.
- W4385225652 hasConcept C114614502 @default.
- W4385225652 hasConcept C121332964 @default.
- W4385225652 hasConcept C138885662 @default.
- W4385225652 hasConcept C153180895 @default.
- W4385225652 hasConcept C154945302 @default.
- W4385225652 hasConcept C163258240 @default.
- W4385225652 hasConcept C178790620 @default.
- W4385225652 hasConcept C184720557 @default.
- W4385225652 hasConcept C185592680 @default.
- W4385225652 hasConcept C2776401178 @default.
- W4385225652 hasConcept C2779227376 @default.
- W4385225652 hasConcept C31258907 @default.
- W4385225652 hasConcept C33676613 @default.
- W4385225652 hasConcept C33923547 @default.
- W4385225652 hasConcept C41008148 @default.
- W4385225652 hasConcept C41895202 @default.
- W4385225652 hasConcept C48145219 @default.
- W4385225652 hasConcept C62520636 @default.
- W4385225652 hasConceptScore W4385225652C11413529 @default.
- W4385225652 hasConceptScore W4385225652C114614502 @default.
- W4385225652 hasConceptScore W4385225652C121332964 @default.
- W4385225652 hasConceptScore W4385225652C138885662 @default.
- W4385225652 hasConceptScore W4385225652C153180895 @default.
- W4385225652 hasConceptScore W4385225652C154945302 @default.
- W4385225652 hasConceptScore W4385225652C163258240 @default.
- W4385225652 hasConceptScore W4385225652C178790620 @default.
- W4385225652 hasConceptScore W4385225652C184720557 @default.
- W4385225652 hasConceptScore W4385225652C185592680 @default.
- W4385225652 hasConceptScore W4385225652C2776401178 @default.
- W4385225652 hasConceptScore W4385225652C2779227376 @default.
- W4385225652 hasConceptScore W4385225652C31258907 @default.
- W4385225652 hasConceptScore W4385225652C33676613 @default.
- W4385225652 hasConceptScore W4385225652C33923547 @default.
- W4385225652 hasConceptScore W4385225652C41008148 @default.
- W4385225652 hasConceptScore W4385225652C41895202 @default.
- W4385225652 hasConceptScore W4385225652C48145219 @default.
- W4385225652 hasConceptScore W4385225652C62520636 @default.
- W4385225652 hasLocation W43852256521 @default.
- W4385225652 hasOpenAccess W4385225652 @default.
- W4385225652 hasPrimaryLocation W43852256521 @default.
- W4385225652 hasRelatedWork W1985412924 @default.
- W4385225652 hasRelatedWork W2033914206 @default.
- W4385225652 hasRelatedWork W2146076056 @default.
- W4385225652 hasRelatedWork W2163831990 @default.
- W4385225652 hasRelatedWork W2375389409 @default.
- W4385225652 hasRelatedWork W2382607599 @default.
- W4385225652 hasRelatedWork W2488051804 @default.
- W4385225652 hasRelatedWork W2546942002 @default.
- W4385225652 hasRelatedWork W3003836766 @default.
- W4385225652 hasRelatedWork W4299878869 @default.
- W4385225652 isParatext "false" @default.
- W4385225652 isRetracted "false" @default.
- W4385225652 workType "article" @default.