Matches in SemOpenAlex for { <https://semopenalex.org/work/W3209306061> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W3209306061 abstract "Adversarial training is originated in image classification to address the problem of adversarial attacks, where an invisible perturbation in an image leads to a significant change in model decision. It recently has been observed to be effective in alleviating the long-tailed classification problem, where an imbalanced size of classes makes the model has much lower performance on small classes. However, existing methods typically focus on the methods to generate perturbations for data, while the contributions of different perturbations to long-tailed classification have not been well analyzed. To this end, this paper presents an investigation on the perturbation generation and incorporation components of existing adversarial training methods and proposes a taxonomy that defines these methods using three levels of components, in terms of information, methodology, and optimization. This taxonomy may serve as a design paradigm where an adversarial training algorithm can be created by combining different components in the taxonomy. A comparative study is conducted to verify the influence of each component in long-tailed classification. Experimental results on two benchmarking datasets show that a combination of statistical perturbations and hybrid optimization achieves a promising performance, and the gradient-based method typically improves the performance of both the head and tail classes. More importantly, it is verified that a reasonable combination of the components in our taxonomy may create an algorithm that outperforms the state-of-the-art." @default.
- W3209306061 created "2021-11-08" @default.
- W3209306061 creator A5006833163 @default.
- W3209306061 creator A5014836905 @default.
- W3209306061 creator A5020274552 @default.
- W3209306061 creator A5029698374 @default.
- W3209306061 date "2021-10-20" @default.
- W3209306061 modified "2023-09-30" @default.
- W3209306061 title "Comparative Study of Adversarial Training Methods for Long-tailed Classification" @default.
- W3209306061 cites W2543927648 @default.
- W3209306061 cites W2962898354 @default.
- W3209306061 cites W2962933664 @default.
- W3209306061 cites W2963351448 @default.
- W3209306061 cites W2963691377 @default.
- W3209306061 cites W2981515171 @default.
- W3209306061 cites W2986674980 @default.
- W3209306061 cites W3010512657 @default.
- W3209306061 cites W3034561829 @default.
- W3209306061 cites W3034601242 @default.
- W3209306061 cites W3096121526 @default.
- W3209306061 cites W3171998492 @default.
- W3209306061 doi "https://doi.org/10.1145/3475724.3483601" @default.
- W3209306061 hasPublicationYear "2021" @default.
- W3209306061 type Work @default.
- W3209306061 sameAs 3209306061 @default.
- W3209306061 citedByCount "7" @default.
- W3209306061 countsByYear W32093060612022 @default.
- W3209306061 countsByYear W32093060612023 @default.
- W3209306061 crossrefType "proceedings-article" @default.
- W3209306061 hasAuthorship W3209306061A5006833163 @default.
- W3209306061 hasAuthorship W3209306061A5014836905 @default.
- W3209306061 hasAuthorship W3209306061A5020274552 @default.
- W3209306061 hasAuthorship W3209306061A5029698374 @default.
- W3209306061 hasConcept C115961682 @default.
- W3209306061 hasConcept C119857082 @default.
- W3209306061 hasConcept C124101348 @default.
- W3209306061 hasConcept C144133560 @default.
- W3209306061 hasConcept C153180895 @default.
- W3209306061 hasConcept C154945302 @default.
- W3209306061 hasConcept C162853370 @default.
- W3209306061 hasConcept C37736160 @default.
- W3209306061 hasConcept C41008148 @default.
- W3209306061 hasConcept C51632099 @default.
- W3209306061 hasConcept C58642233 @default.
- W3209306061 hasConcept C59822182 @default.
- W3209306061 hasConcept C75294576 @default.
- W3209306061 hasConcept C86251818 @default.
- W3209306061 hasConcept C86803240 @default.
- W3209306061 hasConceptScore W3209306061C115961682 @default.
- W3209306061 hasConceptScore W3209306061C119857082 @default.
- W3209306061 hasConceptScore W3209306061C124101348 @default.
- W3209306061 hasConceptScore W3209306061C144133560 @default.
- W3209306061 hasConceptScore W3209306061C153180895 @default.
- W3209306061 hasConceptScore W3209306061C154945302 @default.
- W3209306061 hasConceptScore W3209306061C162853370 @default.
- W3209306061 hasConceptScore W3209306061C37736160 @default.
- W3209306061 hasConceptScore W3209306061C41008148 @default.
- W3209306061 hasConceptScore W3209306061C51632099 @default.
- W3209306061 hasConceptScore W3209306061C58642233 @default.
- W3209306061 hasConceptScore W3209306061C59822182 @default.
- W3209306061 hasConceptScore W3209306061C75294576 @default.
- W3209306061 hasConceptScore W3209306061C86251818 @default.
- W3209306061 hasConceptScore W3209306061C86803240 @default.
- W3209306061 hasFunder F4320321001 @default.
- W3209306061 hasLocation W32093060611 @default.
- W3209306061 hasOpenAccess W3209306061 @default.
- W3209306061 hasPrimaryLocation W32093060611 @default.
- W3209306061 hasRelatedWork W133358225 @default.
- W3209306061 hasRelatedWork W1577137544 @default.
- W3209306061 hasRelatedWork W2508908072 @default.
- W3209306061 hasRelatedWork W2509146328 @default.
- W3209306061 hasRelatedWork W2742991909 @default.
- W3209306061 hasRelatedWork W2996038082 @default.
- W3209306061 hasRelatedWork W3046843850 @default.
- W3209306061 hasRelatedWork W3198701625 @default.
- W3209306061 hasRelatedWork W3201905201 @default.
- W3209306061 hasRelatedWork W4312741812 @default.
- W3209306061 isParatext "false" @default.
- W3209306061 isRetracted "false" @default.
- W3209306061 magId "3209306061" @default.
- W3209306061 workType "article" @default.