Matches in SemOpenAlex for { <https://semopenalex.org/work/W3170571056> ?p ?o ?g. }
Showing items 1 to 78 of
78
with 100 items per page.
- W3170571056 endingPage "8775" @default.
- W3170571056 startingPage "8764" @default.
- W3170571056 abstract "Detecting anomalous inputs, such as adversarial and out-of-distribution (OOD) inputs, is critical for classifiers (including deep neural networks or DNNs) deployed in real-world applications. While prior works have proposed various methods to detect such anomalous samples using information from the internal layer representations of a DNN, there is a lack of consensus on a principled approach for the different components of such a detection method. As a result, often heuristic and one-off methods are applied for different aspects of this problem. We propose an unsupervised anomaly detection framework based on the internal DNN layer representations in the form of a meta-algorithm with configurable components. We proceed to propose specific instantiations for each component of the meta-algorithm based on ideas grounded in statistical testing and anomaly detection. We evaluate the proposed methods on well-known image classification datasets with strong adversarial attacks and OOD inputs, including an adaptive attack that uses the internal layer representations of the DNN (often not considered in prior work). Comparisons with five recently-proposed competing detection methods demonstrates the effectiveness of our method in detecting adversarial and OOD inputs." @default.
- W3170571056 created "2021-06-22" @default.
- W3170571056 creator A5005867534 @default.
- W3170571056 creator A5009092703 @default.
- W3170571056 creator A5033218913 @default.
- W3170571056 creator A5088826068 @default.
- W3170571056 date "2021-07-18" @default.
- W3170571056 modified "2023-09-23" @default.
- W3170571056 title "A General Framework For Detecting Anomalous Inputs to DNN Classifiers" @default.
- W3170571056 hasPublicationYear "2021" @default.
- W3170571056 type Work @default.
- W3170571056 sameAs 3170571056 @default.
- W3170571056 citedByCount "2" @default.
- W3170571056 countsByYear W31705710562021 @default.
- W3170571056 crossrefType "proceedings-article" @default.
- W3170571056 hasAuthorship W3170571056A5005867534 @default.
- W3170571056 hasAuthorship W3170571056A5009092703 @default.
- W3170571056 hasAuthorship W3170571056A5033218913 @default.
- W3170571056 hasAuthorship W3170571056A5088826068 @default.
- W3170571056 hasConcept C119857082 @default.
- W3170571056 hasConcept C121332964 @default.
- W3170571056 hasConcept C153180895 @default.
- W3170571056 hasConcept C154945302 @default.
- W3170571056 hasConcept C168167062 @default.
- W3170571056 hasConcept C173801870 @default.
- W3170571056 hasConcept C178790620 @default.
- W3170571056 hasConcept C185592680 @default.
- W3170571056 hasConcept C2779227376 @default.
- W3170571056 hasConcept C2984842247 @default.
- W3170571056 hasConcept C37736160 @default.
- W3170571056 hasConcept C41008148 @default.
- W3170571056 hasConcept C50644808 @default.
- W3170571056 hasConcept C739882 @default.
- W3170571056 hasConcept C97355855 @default.
- W3170571056 hasConceptScore W3170571056C119857082 @default.
- W3170571056 hasConceptScore W3170571056C121332964 @default.
- W3170571056 hasConceptScore W3170571056C153180895 @default.
- W3170571056 hasConceptScore W3170571056C154945302 @default.
- W3170571056 hasConceptScore W3170571056C168167062 @default.
- W3170571056 hasConceptScore W3170571056C173801870 @default.
- W3170571056 hasConceptScore W3170571056C178790620 @default.
- W3170571056 hasConceptScore W3170571056C185592680 @default.
- W3170571056 hasConceptScore W3170571056C2779227376 @default.
- W3170571056 hasConceptScore W3170571056C2984842247 @default.
- W3170571056 hasConceptScore W3170571056C37736160 @default.
- W3170571056 hasConceptScore W3170571056C41008148 @default.
- W3170571056 hasConceptScore W3170571056C50644808 @default.
- W3170571056 hasConceptScore W3170571056C739882 @default.
- W3170571056 hasConceptScore W3170571056C97355855 @default.
- W3170571056 hasLocation W31705710561 @default.
- W3170571056 hasOpenAccess W3170571056 @default.
- W3170571056 hasPrimaryLocation W31705710561 @default.
- W3170571056 hasRelatedWork W2732428493 @default.
- W3170571056 hasRelatedWork W2907957312 @default.
- W3170571056 hasRelatedWork W2963307331 @default.
- W3170571056 hasRelatedWork W2966734812 @default.
- W3170571056 hasRelatedWork W2968940383 @default.
- W3170571056 hasRelatedWork W2969337066 @default.
- W3170571056 hasRelatedWork W3012922053 @default.
- W3170571056 hasRelatedWork W3026705323 @default.
- W3170571056 hasRelatedWork W3040865353 @default.
- W3170571056 hasRelatedWork W3046973038 @default.
- W3170571056 hasRelatedWork W3081406936 @default.
- W3170571056 hasRelatedWork W3089397955 @default.
- W3170571056 hasRelatedWork W3090897510 @default.
- W3170571056 hasRelatedWork W3100726446 @default.
- W3170571056 hasRelatedWork W3109541357 @default.
- W3170571056 hasRelatedWork W3131088496 @default.
- W3170571056 hasRelatedWork W3136173905 @default.
- W3170571056 hasRelatedWork W3136401541 @default.
- W3170571056 hasRelatedWork W3170633914 @default.
- W3170571056 hasRelatedWork W3197649772 @default.
- W3170571056 isParatext "false" @default.
- W3170571056 isRetracted "false" @default.
- W3170571056 magId "3170571056" @default.
- W3170571056 workType "article" @default.