Matches in SemOpenAlex for { <https://semopenalex.org/work/W2987686867> ?p ?o ?g. }
- W2987686867 abstract "AI Safety is a major concern in many deep learning applications such as autonomous driving. Given a trained deep learning model, an important natural problem is how to reliably verify the model's prediction. In this paper, we propose a novel framework -- deep verifier networks (DVN) to verify the inputs and outputs of deep discriminative models with deep generative models. Our proposed model is based on conditional variational auto-encoders with disentanglement constraints. We give both intuitive and theoretical justifications of the model. Our verifier network is trained independently with the prediction model, which eliminates the need of retraining the verifier network for a new model. We test the verifier network on out-of-distribution detection and adversarial example detection problems, as well as anomaly detection problems in structured prediction tasks such as image caption generation. We achieve state-of-the-art results in all of these problems." @default.
- W2987686867 created "2019-11-22" @default.
- W2987686867 creator A5000578374 @default.
- W2987686867 creator A5026108994 @default.
- W2987686867 creator A5032046813 @default.
- W2987686867 creator A5057271863 @default.
- W2987686867 creator A5085080733 @default.
- W2987686867 creator A5086198262 @default.
- W2987686867 creator A5088216458 @default.
- W2987686867 date "2019-11-17" @default.
- W2987686867 modified "2023-09-24" @default.
- W2987686867 title "Deep Verifier Networks: Verification of Deep Discriminative Models with Deep Generative Models" @default.
- W2987686867 cites W1514535095 @default.
- W2987686867 cites W1779483307 @default.
- W2987686867 cites W1924770834 @default.
- W2987686867 cites W2099471712 @default.
- W2987686867 cites W2108598243 @default.
- W2987686867 cites W2115627867 @default.
- W2987686867 cites W2163605009 @default.
- W2987686867 cites W2210838531 @default.
- W2987686867 cites W2243397390 @default.
- W2987686867 cites W2335728318 @default.
- W2987686867 cites W2398118205 @default.
- W2987686867 cites W2401231614 @default.
- W2987686867 cites W2460937040 @default.
- W2987686867 cites W2462906003 @default.
- W2987686867 cites W2511730936 @default.
- W2987686867 cites W2531327146 @default.
- W2987686867 cites W2557283755 @default.
- W2987686867 cites W2577946330 @default.
- W2987686867 cites W2594867206 @default.
- W2987686867 cites W2601315196 @default.
- W2987686867 cites W2737398044 @default.
- W2987686867 cites W2786712888 @default.
- W2987686867 cites W2793613699 @default.
- W2987686867 cites W2867167548 @default.
- W2987686867 cites W2889625178 @default.
- W2987686867 cites W2890591829 @default.
- W2987686867 cites W2898795295 @default.
- W2987686867 cites W2902453469 @default.
- W2987686867 cites W2903204734 @default.
- W2987686867 cites W2904981516 @default.
- W2987686867 cites W2907511876 @default.
- W2987686867 cites W2913282330 @default.
- W2987686867 cites W2913300775 @default.
- W2987686867 cites W2948571844 @default.
- W2987686867 cites W2948801556 @default.
- W2987686867 cites W2949650786 @default.
- W2987686867 cites W2949999304 @default.
- W2987686867 cites W2950178297 @default.
- W2987686867 cites W2951873722 @default.
- W2987686867 cites W2952816888 @default.
- W2987686867 cites W2963139417 @default.
- W2987686867 cites W2963158386 @default.
- W2987686867 cites W2963207607 @default.
- W2987686867 cites W2963564844 @default.
- W2987686867 cites W2963636093 @default.
- W2987686867 cites W2963693742 @default.
- W2987686867 cites W2964184973 @default.
- W2987686867 cites W2964212410 @default.
- W2987686867 cites W2970602229 @default.
- W2987686867 cites W2970859221 @default.
- W2987686867 cites W2980402966 @default.
- W2987686867 cites W2989128580 @default.
- W2987686867 cites W2993534970 @default.
- W2987686867 cites W2998414876 @default.
- W2987686867 cites W2998802446 @default.
- W2987686867 cites W3002155012 @default.
- W2987686867 cites W3034505102 @default.
- W2987686867 cites W3034566037 @default.
- W2987686867 cites W3035264762 @default.
- W2987686867 cites W3035416875 @default.
- W2987686867 cites W3035831564 @default.
- W2987686867 cites W3042790181 @default.
- W2987686867 cites W3084765493 @default.
- W2987686867 cites W3118182541 @default.
- W2987686867 cites W3118608800 @default.
- W2987686867 cites W3142208146 @default.
- W2987686867 cites W967544008 @default.
- W2987686867 doi "https://doi.org/10.48550/arxiv.1911.07421" @default.
- W2987686867 hasPublicationYear "2019" @default.
- W2987686867 type Work @default.
- W2987686867 sameAs 2987686867 @default.
- W2987686867 citedByCount "11" @default.
- W2987686867 countsByYear W29876868672019 @default.
- W2987686867 countsByYear W29876868672020 @default.
- W2987686867 countsByYear W29876868672021 @default.
- W2987686867 crossrefType "posted-content" @default.
- W2987686867 hasAuthorship W2987686867A5000578374 @default.
- W2987686867 hasAuthorship W2987686867A5026108994 @default.
- W2987686867 hasAuthorship W2987686867A5032046813 @default.
- W2987686867 hasAuthorship W2987686867A5057271863 @default.
- W2987686867 hasAuthorship W2987686867A5085080733 @default.
- W2987686867 hasAuthorship W2987686867A5086198262 @default.
- W2987686867 hasAuthorship W2987686867A5088216458 @default.
- W2987686867 hasBestOaLocation W29876868671 @default.
- W2987686867 hasConcept C101738243 @default.
- W2987686867 hasConcept C108583219 @default.
- W2987686867 hasConcept C111919701 @default.
- W2987686867 hasConcept C118505674 @default.