Matches in SemOpenAlex for { <https://semopenalex.org/work/W3204920603> ?p ?o ?g. }
Showing items 1 to 94 of
94
with 100 items per page.
- W3204920603 endingPage "6523" @default.
- W3204920603 startingPage "6523" @default.
- W3204920603 abstract "Deep neural networks have achieved state-of-the-art performance in image classification. Due to this success, deep learning is now also being applied to other data modalities such as multispectral images, lidar and radar data. However, successfully training a deep neural network requires a large reddataset. Therefore, transitioning to a new sensor modality (e.g., from regular camera images to multispectral camera images) might result in a drop in performance, due to the limited availability of data in the new modality. This might hinder the adoption rate and time to market for new sensor technologies. In this paper, we present an approach to leverage the knowledge of a teacher network, that was trained using the original data modality, to improve the performance of a student network on a new data modality: a technique known in literature as knowledge distillation. By applying knowledge distillation to the problem of sensor transition, we can greatly speed up this process. We validate this approach using a multimodal version of the MNIST dataset. Especially when little data is available in the new modality (i.e., 10 images), training with additional teacher supervision results in increased performance, with the student network scoring a test set accuracy of 0.77, compared to an accuracy of 0.37 for the baseline. We also explore two extensions to the default method of knowledge distillation, which we evaluate on a multimodal version of the CIFAR-10 dataset: an annealing scheme for the hyperparameter α and selective knowledge distillation. Of these two, the first yields the best results. Choosing the optimal annealing scheme results in an increase in test set accuracy of 6%. Finally, we apply our method to the real-world use case of skin lesion classification." @default.
- W3204920603 created "2021-10-11" @default.
- W3204920603 creator A5001314520 @default.
- W3204920603 creator A5032558375 @default.
- W3204920603 creator A5037337384 @default.
- W3204920603 creator A5048747366 @default.
- W3204920603 creator A5053913176 @default.
- W3204920603 creator A5056570824 @default.
- W3204920603 creator A5061031206 @default.
- W3204920603 creator A5071221406 @default.
- W3204920603 date "2021-09-29" @default.
- W3204920603 modified "2023-10-17" @default.
- W3204920603 title "Data-Efficient Sensor Upgrade Path Using Knowledge Distillation" @default.
- W3204920603 cites W2112796928 @default.
- W3204920603 cites W2117539524 @default.
- W3204920603 cites W2166664821 @default.
- W3204920603 cites W2581082771 @default.
- W3204920603 cites W2618530766 @default.
- W3204920603 cites W2757940437 @default.
- W3204920603 cites W2959684014 @default.
- W3204920603 cites W3014066178 @default.
- W3204920603 cites W3102785203 @default.
- W3204920603 doi "https://doi.org/10.3390/s21196523" @default.
- W3204920603 hasPubMedCentralId "https://www.ncbi.nlm.nih.gov/pmc/articles/8512581" @default.
- W3204920603 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/34640843" @default.
- W3204920603 hasPublicationYear "2021" @default.
- W3204920603 type Work @default.
- W3204920603 sameAs 3204920603 @default.
- W3204920603 citedByCount "4" @default.
- W3204920603 countsByYear W32049206032022 @default.
- W3204920603 countsByYear W32049206032023 @default.
- W3204920603 crossrefType "journal-article" @default.
- W3204920603 hasAuthorship W3204920603A5001314520 @default.
- W3204920603 hasAuthorship W3204920603A5032558375 @default.
- W3204920603 hasAuthorship W3204920603A5037337384 @default.
- W3204920603 hasAuthorship W3204920603A5048747366 @default.
- W3204920603 hasAuthorship W3204920603A5053913176 @default.
- W3204920603 hasAuthorship W3204920603A5056570824 @default.
- W3204920603 hasAuthorship W3204920603A5061031206 @default.
- W3204920603 hasAuthorship W3204920603A5071221406 @default.
- W3204920603 hasBestOaLocation W32049206031 @default.
- W3204920603 hasConcept C108583219 @default.
- W3204920603 hasConcept C119857082 @default.
- W3204920603 hasConcept C124101348 @default.
- W3204920603 hasConcept C153083717 @default.
- W3204920603 hasConcept C154945302 @default.
- W3204920603 hasConcept C173163844 @default.
- W3204920603 hasConcept C178790620 @default.
- W3204920603 hasConcept C185592680 @default.
- W3204920603 hasConcept C204030448 @default.
- W3204920603 hasConcept C2780226545 @default.
- W3204920603 hasConcept C41008148 @default.
- W3204920603 hasConcept C50644808 @default.
- W3204920603 hasConcept C81363708 @default.
- W3204920603 hasConceptScore W3204920603C108583219 @default.
- W3204920603 hasConceptScore W3204920603C119857082 @default.
- W3204920603 hasConceptScore W3204920603C124101348 @default.
- W3204920603 hasConceptScore W3204920603C153083717 @default.
- W3204920603 hasConceptScore W3204920603C154945302 @default.
- W3204920603 hasConceptScore W3204920603C173163844 @default.
- W3204920603 hasConceptScore W3204920603C178790620 @default.
- W3204920603 hasConceptScore W3204920603C185592680 @default.
- W3204920603 hasConceptScore W3204920603C204030448 @default.
- W3204920603 hasConceptScore W3204920603C2780226545 @default.
- W3204920603 hasConceptScore W3204920603C41008148 @default.
- W3204920603 hasConceptScore W3204920603C50644808 @default.
- W3204920603 hasConceptScore W3204920603C81363708 @default.
- W3204920603 hasFunder F4320313460 @default.
- W3204920603 hasIssue "19" @default.
- W3204920603 hasLocation W32049206031 @default.
- W3204920603 hasLocation W32049206032 @default.
- W3204920603 hasLocation W32049206033 @default.
- W3204920603 hasLocation W32049206034 @default.
- W3204920603 hasLocation W32049206035 @default.
- W3204920603 hasLocation W32049206036 @default.
- W3204920603 hasOpenAccess W3204920603 @default.
- W3204920603 hasPrimaryLocation W32049206031 @default.
- W3204920603 hasRelatedWork W1511643196 @default.
- W3204920603 hasRelatedWork W1544811710 @default.
- W3204920603 hasRelatedWork W172072032 @default.
- W3204920603 hasRelatedWork W2006066416 @default.
- W3204920603 hasRelatedWork W2032891171 @default.
- W3204920603 hasRelatedWork W2124951708 @default.
- W3204920603 hasRelatedWork W2771047279 @default.
- W3204920603 hasRelatedWork W3157073418 @default.
- W3204920603 hasRelatedWork W4318664220 @default.
- W3204920603 hasRelatedWork W62810557 @default.
- W3204920603 hasVolume "21" @default.
- W3204920603 isParatext "false" @default.
- W3204920603 isRetracted "false" @default.
- W3204920603 magId "3204920603" @default.
- W3204920603 workType "article" @default.