Matches in SemOpenAlex for { <https://semopenalex.org/work/W3109356881> ?p ?o ?g. }
Showing items 1 to 70 of
70
with 100 items per page.
- W3109356881 abstract "Adversarial example generation methods in NLP rely on models like language models or sentence encoders to determine if potential adversarial examples are valid. In these methods, a valid adversarial example fools the model being attacked, and is determined to be semantically or syntactically valid by a second model. Research to date has counted all such examples as errors by the attacked model. We contend that these adversarial examples may not be flaws in the attacked model, but flaws in the model that determines validity. We term such invalid inputs second-order adversarial examples. We propose the constraint robustness curve and associated metric ACCS as tools for evaluating the robustness of a constraint to second-order adversarial examples. To generate this curve, we design an adversarial attack to run directly on the semantic similarity models. We test on two constraints, the Universal Sentence Encoder (USE) and BERTScore. Our findings indicate that such second-order examples exist, but are typically less common than first-order adversarial examples in state-of-the-art models. They also indicate that USE is effective as constraint on NLP adversarial examples, while BERTScore is nearly ineffectual. Code for running the experiments in this paper is available at this https URL." @default.
- W3109356881 created "2020-12-07" @default.
- W3109356881 creator A5005724403 @default.
- W3109356881 date "2020-01-01" @default.
- W3109356881 modified "2023-10-16" @default.
- W3109356881 title "Second-Order NLP Adversarial Examples" @default.
- W3109356881 doi "https://doi.org/10.18653/v1/2020.blackboxnlp-1.22" @default.
- W3109356881 hasPublicationYear "2020" @default.
- W3109356881 type Work @default.
- W3109356881 sameAs 3109356881 @default.
- W3109356881 citedByCount "0" @default.
- W3109356881 crossrefType "proceedings-article" @default.
- W3109356881 hasAuthorship W3109356881A5005724403 @default.
- W3109356881 hasBestOaLocation W31093568811 @default.
- W3109356881 hasConcept C104317684 @default.
- W3109356881 hasConcept C111919701 @default.
- W3109356881 hasConcept C118505674 @default.
- W3109356881 hasConcept C119857082 @default.
- W3109356881 hasConcept C137293760 @default.
- W3109356881 hasConcept C154945302 @default.
- W3109356881 hasConcept C162324750 @default.
- W3109356881 hasConcept C176217482 @default.
- W3109356881 hasConcept C185592680 @default.
- W3109356881 hasConcept C204321447 @default.
- W3109356881 hasConcept C21547014 @default.
- W3109356881 hasConcept C2524010 @default.
- W3109356881 hasConcept C2776036281 @default.
- W3109356881 hasConcept C2777530160 @default.
- W3109356881 hasConcept C33923547 @default.
- W3109356881 hasConcept C37736160 @default.
- W3109356881 hasConcept C41008148 @default.
- W3109356881 hasConcept C55493867 @default.
- W3109356881 hasConcept C63479239 @default.
- W3109356881 hasConceptScore W3109356881C104317684 @default.
- W3109356881 hasConceptScore W3109356881C111919701 @default.
- W3109356881 hasConceptScore W3109356881C118505674 @default.
- W3109356881 hasConceptScore W3109356881C119857082 @default.
- W3109356881 hasConceptScore W3109356881C137293760 @default.
- W3109356881 hasConceptScore W3109356881C154945302 @default.
- W3109356881 hasConceptScore W3109356881C162324750 @default.
- W3109356881 hasConceptScore W3109356881C176217482 @default.
- W3109356881 hasConceptScore W3109356881C185592680 @default.
- W3109356881 hasConceptScore W3109356881C204321447 @default.
- W3109356881 hasConceptScore W3109356881C21547014 @default.
- W3109356881 hasConceptScore W3109356881C2524010 @default.
- W3109356881 hasConceptScore W3109356881C2776036281 @default.
- W3109356881 hasConceptScore W3109356881C2777530160 @default.
- W3109356881 hasConceptScore W3109356881C33923547 @default.
- W3109356881 hasConceptScore W3109356881C37736160 @default.
- W3109356881 hasConceptScore W3109356881C41008148 @default.
- W3109356881 hasConceptScore W3109356881C55493867 @default.
- W3109356881 hasConceptScore W3109356881C63479239 @default.
- W3109356881 hasLocation W31093568811 @default.
- W3109356881 hasLocation W31093568812 @default.
- W3109356881 hasOpenAccess W3109356881 @default.
- W3109356881 hasPrimaryLocation W31093568811 @default.
- W3109356881 hasRelatedWork W12215899 @default.
- W3109356881 hasRelatedWork W12732426 @default.
- W3109356881 hasRelatedWork W17684 @default.
- W3109356881 hasRelatedWork W4538592 @default.
- W3109356881 hasRelatedWork W7619760 @default.
- W3109356881 hasRelatedWork W8053366 @default.
- W3109356881 hasRelatedWork W90294 @default.
- W3109356881 hasRelatedWork W9218159 @default.
- W3109356881 hasRelatedWork W9840409 @default.
- W3109356881 hasRelatedWork W13002482 @default.
- W3109356881 isParatext "false" @default.
- W3109356881 isRetracted "false" @default.
- W3109356881 magId "3109356881" @default.
- W3109356881 workType "article" @default.