Matches in SemOpenAlex for { <https://semopenalex.org/work/W3037045905> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W3037045905 endingPage "8391" @default.
- W3037045905 startingPage "8384" @default.
- W3037045905 abstract "Text classification is a basic task in natural language processing, but the small character perturbations in words can greatly decrease the effectiveness of text classification models, which is called character-level adversarial example attack. There are two main challenges in character-level adversarial examples defense, which are out-of-vocabulary words in word embedding model and the distribution difference between training and inference. Both of these two challenges make the character-level adversarial examples difficult to defend. In this paper, we propose a framework which jointly uses the character embedding and the adversarial stability training to overcome these two challenges. Our experimental results on five text classification data sets show that the models based on our framework can effectively defend character-level adversarial examples, and our models can defend 93.19% gradient-based adversarial examples and 94.83% natural adversarial examples, which outperforms the state-of-the-art defense models." @default.
- W3037045905 created "2020-07-02" @default.
- W3037045905 creator A5040565656 @default.
- W3037045905 creator A5044301848 @default.
- W3037045905 creator A5045460876 @default.
- W3037045905 creator A5057024795 @default.
- W3037045905 creator A5067997634 @default.
- W3037045905 date "2020-04-03" @default.
- W3037045905 modified "2023-10-16" @default.
- W3037045905 title "Joint Character-Level Word Embedding and Adversarial Stability Training to Defend Adversarial Text" @default.
- W3037045905 doi "https://doi.org/10.1609/aaai.v34i05.6356" @default.
- W3037045905 hasPublicationYear "2020" @default.
- W3037045905 type Work @default.
- W3037045905 sameAs 3037045905 @default.
- W3037045905 citedByCount "18" @default.
- W3037045905 countsByYear W30370459052020 @default.
- W3037045905 countsByYear W30370459052021 @default.
- W3037045905 countsByYear W30370459052022 @default.
- W3037045905 countsByYear W30370459052023 @default.
- W3037045905 crossrefType "journal-article" @default.
- W3037045905 hasAuthorship W3037045905A5040565656 @default.
- W3037045905 hasAuthorship W3037045905A5044301848 @default.
- W3037045905 hasAuthorship W3037045905A5045460876 @default.
- W3037045905 hasAuthorship W3037045905A5057024795 @default.
- W3037045905 hasAuthorship W3037045905A5067997634 @default.
- W3037045905 hasBestOaLocation W30370459051 @default.
- W3037045905 hasConcept C112972136 @default.
- W3037045905 hasConcept C119857082 @default.
- W3037045905 hasConcept C127413603 @default.
- W3037045905 hasConcept C138885662 @default.
- W3037045905 hasConcept C154945302 @default.
- W3037045905 hasConcept C201995342 @default.
- W3037045905 hasConcept C204321447 @default.
- W3037045905 hasConcept C2524010 @default.
- W3037045905 hasConcept C2777601683 @default.
- W3037045905 hasConcept C2780451532 @default.
- W3037045905 hasConcept C2780861071 @default.
- W3037045905 hasConcept C33923547 @default.
- W3037045905 hasConcept C37736160 @default.
- W3037045905 hasConcept C41008148 @default.
- W3037045905 hasConcept C41608201 @default.
- W3037045905 hasConcept C41895202 @default.
- W3037045905 hasConcept C90805587 @default.
- W3037045905 hasConceptScore W3037045905C112972136 @default.
- W3037045905 hasConceptScore W3037045905C119857082 @default.
- W3037045905 hasConceptScore W3037045905C127413603 @default.
- W3037045905 hasConceptScore W3037045905C138885662 @default.
- W3037045905 hasConceptScore W3037045905C154945302 @default.
- W3037045905 hasConceptScore W3037045905C201995342 @default.
- W3037045905 hasConceptScore W3037045905C204321447 @default.
- W3037045905 hasConceptScore W3037045905C2524010 @default.
- W3037045905 hasConceptScore W3037045905C2777601683 @default.
- W3037045905 hasConceptScore W3037045905C2780451532 @default.
- W3037045905 hasConceptScore W3037045905C2780861071 @default.
- W3037045905 hasConceptScore W3037045905C33923547 @default.
- W3037045905 hasConceptScore W3037045905C37736160 @default.
- W3037045905 hasConceptScore W3037045905C41008148 @default.
- W3037045905 hasConceptScore W3037045905C41608201 @default.
- W3037045905 hasConceptScore W3037045905C41895202 @default.
- W3037045905 hasConceptScore W3037045905C90805587 @default.
- W3037045905 hasIssue "05" @default.
- W3037045905 hasLocation W30370459051 @default.
- W3037045905 hasOpenAccess W3037045905 @default.
- W3037045905 hasPrimaryLocation W30370459051 @default.
- W3037045905 hasRelatedWork W1539050421 @default.
- W3037045905 hasRelatedWork W2081647779 @default.
- W3037045905 hasRelatedWork W2363407155 @default.
- W3037045905 hasRelatedWork W2804608824 @default.
- W3037045905 hasRelatedWork W2988804282 @default.
- W3037045905 hasRelatedWork W3137145434 @default.
- W3037045905 hasRelatedWork W3185852197 @default.
- W3037045905 hasRelatedWork W4221161337 @default.
- W3037045905 hasRelatedWork W4301110457 @default.
- W3037045905 hasRelatedWork W77769009 @default.
- W3037045905 hasVolume "34" @default.
- W3037045905 isParatext "false" @default.
- W3037045905 isRetracted "false" @default.
- W3037045905 magId "3037045905" @default.
- W3037045905 workType "article" @default.