Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387010624> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W4387010624 endingPage "106298" @default.
- W4387010624 startingPage "106289" @default.
- W4387010624 abstract "Network Augmentation (NetAug) is a recent method used to improve the performance of tiny neural networks on large-scale datasets. This method provides additional supervision to tiny models from larger augmented models, mitigating the issue of underfitting. However, the capacity of the augmented models is not fully utilized, resulting in underutilization of resources. In order to fully utilize the capacity of a larger augmented model without exacerbating the underfitting of a tiny model, we propose a new method called Multi-Input Network Augmentation (MINA). MINA converts tiny neural networks into a multi-input configuration, allowing only the augmented model to receive more diverse inputs during training. Additionally, tiny neural network can be converted back into their original single-input configuration after training. Our extensive experiments on large-scale datasets demonstrate that MINA is effective in improving the performance of tiny neural networks. We also demonstrate that MINA is consistently effective in downstream tasks, such as fine-grained image classification tasks and object detection tasks." @default.
- W4387010624 created "2023-09-26" @default.
- W4387010624 creator A5020493185 @default.
- W4387010624 creator A5068058738 @default.
- W4387010624 creator A5069036873 @default.
- W4387010624 date "2023-01-01" @default.
- W4387010624 modified "2023-10-09" @default.
- W4387010624 title "MINA: Multi-Input Network Augmentation for Enhancing Tiny Deep Learning" @default.
- W4387010624 cites W12634471 @default.
- W4387010624 cites W1977295328 @default.
- W4387010624 cites W2031489346 @default.
- W4387010624 cites W2108598243 @default.
- W4387010624 cites W2117539524 @default.
- W4387010624 cites W2138011018 @default.
- W4387010624 cites W2194775991 @default.
- W4387010624 cites W2412782625 @default.
- W4387010624 cites W2533598788 @default.
- W4387010624 cites W2949736877 @default.
- W4387010624 cites W2963163009 @default.
- W4387010624 cites W2963351448 @default.
- W4387010624 cites W2963363373 @default.
- W4387010624 cites W2963855133 @default.
- W4387010624 cites W2982083293 @default.
- W4387010624 cites W2984016429 @default.
- W4387010624 cites W2992308087 @default.
- W4387010624 cites W3034429256 @default.
- W4387010624 cites W3035452548 @default.
- W4387010624 cites W3035682985 @default.
- W4387010624 cites W3167976421 @default.
- W4387010624 cites W3171038842 @default.
- W4387010624 cites W4312363311 @default.
- W4387010624 cites W4312662180 @default.
- W4387010624 doi "https://doi.org/10.1109/access.2023.3319313" @default.
- W4387010624 hasPublicationYear "2023" @default.
- W4387010624 type Work @default.
- W4387010624 citedByCount "0" @default.
- W4387010624 crossrefType "journal-article" @default.
- W4387010624 hasAuthorship W4387010624A5020493185 @default.
- W4387010624 hasAuthorship W4387010624A5068058738 @default.
- W4387010624 hasAuthorship W4387010624A5069036873 @default.
- W4387010624 hasBestOaLocation W43870106241 @default.
- W4387010624 hasConcept C108583219 @default.
- W4387010624 hasConcept C119857082 @default.
- W4387010624 hasConcept C121332964 @default.
- W4387010624 hasConcept C154945302 @default.
- W4387010624 hasConcept C2778755073 @default.
- W4387010624 hasConcept C2781238097 @default.
- W4387010624 hasConcept C2984842247 @default.
- W4387010624 hasConcept C41008148 @default.
- W4387010624 hasConcept C50644808 @default.
- W4387010624 hasConcept C62520636 @default.
- W4387010624 hasConceptScore W4387010624C108583219 @default.
- W4387010624 hasConceptScore W4387010624C119857082 @default.
- W4387010624 hasConceptScore W4387010624C121332964 @default.
- W4387010624 hasConceptScore W4387010624C154945302 @default.
- W4387010624 hasConceptScore W4387010624C2778755073 @default.
- W4387010624 hasConceptScore W4387010624C2781238097 @default.
- W4387010624 hasConceptScore W4387010624C2984842247 @default.
- W4387010624 hasConceptScore W4387010624C41008148 @default.
- W4387010624 hasConceptScore W4387010624C50644808 @default.
- W4387010624 hasConceptScore W4387010624C62520636 @default.
- W4387010624 hasFunder F4320322030 @default.
- W4387010624 hasFunder F4320322120 @default.
- W4387010624 hasLocation W43870106241 @default.
- W4387010624 hasOpenAccess W4387010624 @default.
- W4387010624 hasPrimaryLocation W43870106241 @default.
- W4387010624 hasRelatedWork W2611989081 @default.
- W4387010624 hasRelatedWork W2731899572 @default.
- W4387010624 hasRelatedWork W2799384463 @default.
- W4387010624 hasRelatedWork W3000197790 @default.
- W4387010624 hasRelatedWork W3193857078 @default.
- W4387010624 hasRelatedWork W3208304128 @default.
- W4387010624 hasRelatedWork W4230611425 @default.
- W4387010624 hasRelatedWork W4304166257 @default.
- W4387010624 hasRelatedWork W4375867731 @default.
- W4387010624 hasRelatedWork W4377865163 @default.
- W4387010624 hasVolume "11" @default.
- W4387010624 isParatext "false" @default.
- W4387010624 isRetracted "false" @default.
- W4387010624 workType "article" @default.