Matches in SemOpenAlex for { <https://semopenalex.org/work/W4387647987> ?p ?o ?g. }
Showing items 1 to 57 of
57
with 100 items per page.
- W4387647987 abstract "Randomized smoothing has recently attracted attentions in the field of adversarial robustness to provide provable robustness guarantees on smoothed neural network classifiers. However, existing works show that vanilla randomized smoothing usually does not provide good robustness performance and often requires (re)training techniques on the base classifier in order to boost the robustness of the resulting smoothed classifier. In this work, we propose two cost-effective approaches to boost the robustness of randomized smoothing while preserving its clean performance. The first approach introduces a new robust training method AdvMacerwhich combines adversarial training and robustness certification maximization for randomized smoothing. We show that AdvMacer can improve the robustness performance of randomized smoothing classifiers compared to SOTA baselines, while being 3x faster to train than MACER baseline. The second approach introduces a post-processing method EsbRS which greatly improves the robustness certificate based on building model ensembles. We explore different aspects of model ensembles that has not been studied by prior works and propose a novel design methodology to further improve robustness of the ensemble based on our theoretical analysis." @default.
- W4387647987 created "2023-10-15" @default.
- W4387647987 creator A5007306953 @default.
- W4387647987 creator A5030725597 @default.
- W4387647987 creator A5035602232 @default.
- W4387647987 creator A5049043092 @default.
- W4387647987 date "2023-10-11" @default.
- W4387647987 modified "2023-10-15" @default.
- W4387647987 title "Promoting Robustness of Randomized Smoothing: Two Cost-Effective Approaches" @default.
- W4387647987 doi "https://doi.org/10.48550/arxiv.2310.07780" @default.
- W4387647987 hasPublicationYear "2023" @default.
- W4387647987 type Work @default.
- W4387647987 citedByCount "0" @default.
- W4387647987 crossrefType "posted-content" @default.
- W4387647987 hasAuthorship W4387647987A5007306953 @default.
- W4387647987 hasAuthorship W4387647987A5030725597 @default.
- W4387647987 hasAuthorship W4387647987A5035602232 @default.
- W4387647987 hasAuthorship W4387647987A5049043092 @default.
- W4387647987 hasBestOaLocation W43876479871 @default.
- W4387647987 hasConcept C104317684 @default.
- W4387647987 hasConcept C119857082 @default.
- W4387647987 hasConcept C124101348 @default.
- W4387647987 hasConcept C154945302 @default.
- W4387647987 hasConcept C185592680 @default.
- W4387647987 hasConcept C31972630 @default.
- W4387647987 hasConcept C3770464 @default.
- W4387647987 hasConcept C41008148 @default.
- W4387647987 hasConcept C55493867 @default.
- W4387647987 hasConcept C63479239 @default.
- W4387647987 hasConcept C95623464 @default.
- W4387647987 hasConceptScore W4387647987C104317684 @default.
- W4387647987 hasConceptScore W4387647987C119857082 @default.
- W4387647987 hasConceptScore W4387647987C124101348 @default.
- W4387647987 hasConceptScore W4387647987C154945302 @default.
- W4387647987 hasConceptScore W4387647987C185592680 @default.
- W4387647987 hasConceptScore W4387647987C31972630 @default.
- W4387647987 hasConceptScore W4387647987C3770464 @default.
- W4387647987 hasConceptScore W4387647987C41008148 @default.
- W4387647987 hasConceptScore W4387647987C55493867 @default.
- W4387647987 hasConceptScore W4387647987C63479239 @default.
- W4387647987 hasConceptScore W4387647987C95623464 @default.
- W4387647987 hasLocation W43876479871 @default.
- W4387647987 hasOpenAccess W4387647987 @default.
- W4387647987 hasPrimaryLocation W43876479871 @default.
- W4387647987 hasRelatedWork W1978572805 @default.
- W4387647987 hasRelatedWork W1987225439 @default.
- W4387647987 hasRelatedWork W1989941125 @default.
- W4387647987 hasRelatedWork W1997992934 @default.
- W4387647987 hasRelatedWork W2019977573 @default.
- W4387647987 hasRelatedWork W2125114371 @default.
- W4387647987 hasRelatedWork W2149980199 @default.
- W4387647987 hasRelatedWork W2383807498 @default.
- W4387647987 hasRelatedWork W2391139946 @default.
- W4387647987 hasRelatedWork W3125766170 @default.
- W4387647987 isParatext "false" @default.
- W4387647987 isRetracted "false" @default.
- W4387647987 workType "article" @default.