Matches in SemOpenAlex for { <https://semopenalex.org/work/W2807661964> ?p ?o ?g. }
Showing items 1 to 87 of
87
with 100 items per page.
- W2807661964 abstract "Automatic facial expression recognition systems are usually trained from target labels that model each example as belonging unambiguously to a single class (e.g., non-engaged, very engaged, etc.). However, in some settings, ground-truth labels can be more aptly modeled as probability distributions (e.g., [0.1, 0.1, 0.5, 0.3] over 4 engagement categories) that capture the uncertainty that can arise during the annotation process. In this paper, we explore how harnessing the full probability distribution of each label (soft labels), rather than just a scalar summary statistic (hard labels, e.g., majority class or mean), can yield better recognition accuracy when training automated detectors. Our results on a face image dataset (10698 faces over 20 subjects) labeled for perceived student engagement suggest that training on soft labels can deliver engagement detectors that fit the data stat. sig. more accurately (lower cross-entropy for classification, higher Pearson correlation for regression) than when training on hard labels. Moreover, we explore possible reasons for this effect and provide evidence that it is due to implicit regularization that the soft labels enact on the trained engagement detector. This effect is similar to, but empirically seems stronger than, the label smoothing approach proposed by Szegedy, et al. [1]." @default.
- W2807661964 created "2018-06-13" @default.
- W2807661964 creator A5027788582 @default.
- W2807661964 creator A5056976439 @default.
- W2807661964 date "2018-05-01" @default.
- W2807661964 modified "2023-09-24" @default.
- W2807661964 title "Harnessing Label Uncertainty to Improve Modeling: An Application to Student Engagement Recognition" @default.
- W2807661964 cites W1493730813 @default.
- W2807661964 cites W1918539772 @default.
- W2807661964 cites W1966797434 @default.
- W2807661964 cites W1968600824 @default.
- W2807661964 cites W2081112272 @default.
- W2807661964 cites W2087652715 @default.
- W2807661964 cites W2098357182 @default.
- W2807661964 cites W2107671593 @default.
- W2807661964 cites W2108143062 @default.
- W2807661964 cites W2113516703 @default.
- W2807661964 cites W2115417283 @default.
- W2807661964 cites W2148551591 @default.
- W2807661964 cites W2183341477 @default.
- W2807661964 cites W2186046013 @default.
- W2807661964 cites W2345031027 @default.
- W2807661964 cites W2407034139 @default.
- W2807661964 cites W2408871143 @default.
- W2807661964 cites W2485075409 @default.
- W2807661964 cites W4250610610 @default.
- W2807661964 doi "https://doi.org/10.1109/fg.2018.00033" @default.
- W2807661964 hasPublicationYear "2018" @default.
- W2807661964 type Work @default.
- W2807661964 sameAs 2807661964 @default.
- W2807661964 citedByCount "18" @default.
- W2807661964 countsByYear W28076619642018 @default.
- W2807661964 countsByYear W28076619642019 @default.
- W2807661964 countsByYear W28076619642020 @default.
- W2807661964 countsByYear W28076619642021 @default.
- W2807661964 countsByYear W28076619642022 @default.
- W2807661964 countsByYear W28076619642023 @default.
- W2807661964 crossrefType "proceedings-article" @default.
- W2807661964 hasAuthorship W2807661964A5027788582 @default.
- W2807661964 hasAuthorship W2807661964A5056976439 @default.
- W2807661964 hasConcept C105795698 @default.
- W2807661964 hasConcept C106301342 @default.
- W2807661964 hasConcept C119857082 @default.
- W2807661964 hasConcept C121332964 @default.
- W2807661964 hasConcept C153180895 @default.
- W2807661964 hasConcept C154945302 @default.
- W2807661964 hasConcept C2776135515 @default.
- W2807661964 hasConcept C31972630 @default.
- W2807661964 hasConcept C33923547 @default.
- W2807661964 hasConcept C3770464 @default.
- W2807661964 hasConcept C41008148 @default.
- W2807661964 hasConcept C62520636 @default.
- W2807661964 hasConcept C76155785 @default.
- W2807661964 hasConcept C89128539 @default.
- W2807661964 hasConcept C94915269 @default.
- W2807661964 hasConceptScore W2807661964C105795698 @default.
- W2807661964 hasConceptScore W2807661964C106301342 @default.
- W2807661964 hasConceptScore W2807661964C119857082 @default.
- W2807661964 hasConceptScore W2807661964C121332964 @default.
- W2807661964 hasConceptScore W2807661964C153180895 @default.
- W2807661964 hasConceptScore W2807661964C154945302 @default.
- W2807661964 hasConceptScore W2807661964C2776135515 @default.
- W2807661964 hasConceptScore W2807661964C31972630 @default.
- W2807661964 hasConceptScore W2807661964C33923547 @default.
- W2807661964 hasConceptScore W2807661964C3770464 @default.
- W2807661964 hasConceptScore W2807661964C41008148 @default.
- W2807661964 hasConceptScore W2807661964C62520636 @default.
- W2807661964 hasConceptScore W2807661964C76155785 @default.
- W2807661964 hasConceptScore W2807661964C89128539 @default.
- W2807661964 hasConceptScore W2807661964C94915269 @default.
- W2807661964 hasLocation W28076619641 @default.
- W2807661964 hasOpenAccess W2807661964 @default.
- W2807661964 hasPrimaryLocation W28076619641 @default.
- W2807661964 hasRelatedWork W2051752773 @default.
- W2807661964 hasRelatedWork W2079864875 @default.
- W2807661964 hasRelatedWork W2382706727 @default.
- W2807661964 hasRelatedWork W2961085424 @default.
- W2807661964 hasRelatedWork W4213275102 @default.
- W2807661964 hasRelatedWork W4285260836 @default.
- W2807661964 hasRelatedWork W4286629047 @default.
- W2807661964 hasRelatedWork W4306321456 @default.
- W2807661964 hasRelatedWork W4306674287 @default.
- W2807661964 hasRelatedWork W4224009465 @default.
- W2807661964 isParatext "false" @default.
- W2807661964 isRetracted "false" @default.
- W2807661964 magId "2807661964" @default.
- W2807661964 workType "article" @default.