Matches in SemOpenAlex for { <https://semopenalex.org/work/W4385965352> ?p ?o ?g. }
Showing items 1 to 57 of
57
with 100 items per page.
- W4385965352 abstract "Given a classifier, the inherent property of semantic Out-of-Distribution (OOD) samples is that their contents differ from all legal classes in terms of semantics, namely semantic mismatch. There is a recent work that directly applies it to OOD detection, which employs a conditional Generative Adversarial Network (cGAN) to enlarge semantic mismatch in the image space. While achieving remarkable OOD detection performance on small datasets, it is not applicable to ImageNet-scale datasets due to the difficulty in training cGANs with both input images and labels as conditions. As diffusion models are much easier to train and amenable to various conditions compared to cGANs, in this work, we propose to directly use pre-trained diffusion models for semantic mismatch-guided OOD detection, named DiffGuard. Specifically, given an OOD input image and the predicted label from the classifier, we try to enlarge the semantic difference between the reconstructed OOD image under these conditions and the original input image. We also present several test-time techniques to further strengthen such differences. Experimental results show that DiffGuard is effective on both Cifar-10 and hard cases of the large-scale ImageNet, and it can be easily combined with existing OOD detection techniques to achieve state-of-the-art OOD detection results." @default.
- W4385965352 created "2023-08-18" @default.
- W4385965352 creator A5003038962 @default.
- W4385965352 creator A5003882618 @default.
- W4385965352 creator A5010695922 @default.
- W4385965352 creator A5041001060 @default.
- W4385965352 date "2023-08-15" @default.
- W4385965352 modified "2023-09-26" @default.
- W4385965352 title "DiffGuard: Semantic Mismatch-Guided Out-of-Distribution Detection using Pre-trained Diffusion Models" @default.
- W4385965352 doi "https://doi.org/10.48550/arxiv.2308.07687" @default.
- W4385965352 hasPublicationYear "2023" @default.
- W4385965352 type Work @default.
- W4385965352 citedByCount "0" @default.
- W4385965352 crossrefType "posted-content" @default.
- W4385965352 hasAuthorship W4385965352A5003038962 @default.
- W4385965352 hasAuthorship W4385965352A5003882618 @default.
- W4385965352 hasAuthorship W4385965352A5010695922 @default.
- W4385965352 hasAuthorship W4385965352A5041001060 @default.
- W4385965352 hasBestOaLocation W43859653521 @default.
- W4385965352 hasConcept C115961682 @default.
- W4385965352 hasConcept C119857082 @default.
- W4385965352 hasConcept C153180895 @default.
- W4385965352 hasConcept C154945302 @default.
- W4385965352 hasConcept C184337299 @default.
- W4385965352 hasConcept C199360897 @default.
- W4385965352 hasConcept C2986420190 @default.
- W4385965352 hasConcept C2988773926 @default.
- W4385965352 hasConcept C39890363 @default.
- W4385965352 hasConcept C41008148 @default.
- W4385965352 hasConcept C95623464 @default.
- W4385965352 hasConceptScore W4385965352C115961682 @default.
- W4385965352 hasConceptScore W4385965352C119857082 @default.
- W4385965352 hasConceptScore W4385965352C153180895 @default.
- W4385965352 hasConceptScore W4385965352C154945302 @default.
- W4385965352 hasConceptScore W4385965352C184337299 @default.
- W4385965352 hasConceptScore W4385965352C199360897 @default.
- W4385965352 hasConceptScore W4385965352C2986420190 @default.
- W4385965352 hasConceptScore W4385965352C2988773926 @default.
- W4385965352 hasConceptScore W4385965352C39890363 @default.
- W4385965352 hasConceptScore W4385965352C41008148 @default.
- W4385965352 hasConceptScore W4385965352C95623464 @default.
- W4385965352 hasLocation W43859653521 @default.
- W4385965352 hasOpenAccess W4385965352 @default.
- W4385965352 hasPrimaryLocation W43859653521 @default.
- W4385965352 hasRelatedWork W2167582322 @default.
- W4385965352 hasRelatedWork W2563096758 @default.
- W4385965352 hasRelatedWork W2598991778 @default.
- W4385965352 hasRelatedWork W2972035100 @default.
- W4385965352 hasRelatedWork W3139475326 @default.
- W4385965352 hasRelatedWork W3156291593 @default.
- W4385965352 hasRelatedWork W3178813832 @default.
- W4385965352 hasRelatedWork W3198184493 @default.
- W4385965352 hasRelatedWork W3200179079 @default.
- W4385965352 hasRelatedWork W3158004940 @default.
- W4385965352 isParatext "false" @default.
- W4385965352 isRetracted "false" @default.
- W4385965352 workType "article" @default.