Matches in SemOpenAlex for { <https://semopenalex.org/work/W4377693532> ?p ?o ?g. }
Showing items 1 to 48 of
48
with 100 items per page.
- W4377693532 abstract "Out-of-Distribution (OOD) detection is critical for preventing deep learning models from making incorrect predictions, especially in safety-critical applications such as medical diagnosis and autonomous driving. However, neural networks often suffer from overconfidence, making high confidence predictions for OOD data that are never seen during training and may be irrelevant to training data. Determining the reliability of the prediction is still a difficult and challenging task. To address this challenge, we propose a new method called Uncertainty-Estimation with Normalized Logits (UE-NL) for robust learning and OOD detection. The method has three main benefits: (1) Neural networks with UE-NL treat every In Distribution (ID) sample equally by predicting the uncertainty score of input data and the uncertainty is added into SoftMax function to adjust the learning strength of easy and hard samples during training phase, making the model learn robustly and accurately. (2) UE-NL enforces a constant vector norm on the logits to decouple the effect of the increasing output’s norm from optimization process, which causes the overconfidence issue to some extent. (3) UE-NL provides a new metric, the magnitude of uncertainty score, to detect OOD data. Experiments demonstrate that UE-NL outperforms existing methods on common OOD benchmarks and is more robust to noisy ID data that may be misjudged as OOD data by other methods." @default.
- W4377693532 created "2023-05-24" @default.
- W4377693532 creator A5005974219 @default.
- W4377693532 creator A5072861783 @default.
- W4377693532 date "2023-05-23" @default.
- W4377693532 modified "2023-09-26" @default.
- W4377693532 title "Uncertainty-estimation with normalized logits for out-of-distribution detection" @default.
- W4377693532 doi "https://doi.org/10.1117/12.2681144" @default.
- W4377693532 hasPublicationYear "2023" @default.
- W4377693532 type Work @default.
- W4377693532 citedByCount "0" @default.
- W4377693532 crossrefType "proceedings-article" @default.
- W4377693532 hasAuthorship W4377693532A5005974219 @default.
- W4377693532 hasAuthorship W4377693532A5072861783 @default.
- W4377693532 hasBestOaLocation W43776935322 @default.
- W4377693532 hasConcept C105795698 @default.
- W4377693532 hasConcept C110121322 @default.
- W4377693532 hasConcept C127413603 @default.
- W4377693532 hasConcept C134306372 @default.
- W4377693532 hasConcept C201995342 @default.
- W4377693532 hasConcept C33923547 @default.
- W4377693532 hasConcept C41008148 @default.
- W4377693532 hasConcept C96250715 @default.
- W4377693532 hasConceptScore W4377693532C105795698 @default.
- W4377693532 hasConceptScore W4377693532C110121322 @default.
- W4377693532 hasConceptScore W4377693532C127413603 @default.
- W4377693532 hasConceptScore W4377693532C134306372 @default.
- W4377693532 hasConceptScore W4377693532C201995342 @default.
- W4377693532 hasConceptScore W4377693532C33923547 @default.
- W4377693532 hasConceptScore W4377693532C41008148 @default.
- W4377693532 hasConceptScore W4377693532C96250715 @default.
- W4377693532 hasLocation W43776935321 @default.
- W4377693532 hasLocation W43776935322 @default.
- W4377693532 hasOpenAccess W4377693532 @default.
- W4377693532 hasPrimaryLocation W43776935321 @default.
- W4377693532 hasRelatedWork W1525963979 @default.
- W4377693532 hasRelatedWork W2026050033 @default.
- W4377693532 hasRelatedWork W2065867141 @default.
- W4377693532 hasRelatedWork W2099624298 @default.
- W4377693532 hasRelatedWork W228833732 @default.
- W4377693532 hasRelatedWork W2323821083 @default.
- W4377693532 hasRelatedWork W261652165 @default.
- W4377693532 hasRelatedWork W2896041930 @default.
- W4377693532 hasRelatedWork W309072737 @default.
- W4377693532 hasRelatedWork W2183433221 @default.
- W4377693532 isParatext "false" @default.
- W4377693532 isRetracted "false" @default.
- W4377693532 workType "article" @default.