FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4304080660> ?p ?o ?g. }

• W4304080660 abstract "Significant progress has been made in learning image classification neural networks under long-tail data distribution using robust training algorithms such as data re-sampling, re-weighting, and margin adjustment. Those methods, however, ignore the impact of data imbalance on feature normalization. The dominance of majority classes (head classes) in estimating statistics and affine parameters causes internal covariate shifts within less-frequent categories to be overlooked. To alleviate this challenge, we propose a compound batch normalization method based on a Gaussian mixture. It can model the feature space more comprehensively and reduce the dominance of head classes. In addition, a moving average-based expectation maximization (EM) algorithm is employed to estimate the statistical parameters of multiple Gaussian distributions. However, the EM algorithm is sensitive to initialization and can easily become stuck in local minima where the multiple Gaussian components continue to focus on majority classes. To tackle this issue, we developed a dual-path learning framework that employs class-aware split feature normalization to diversify the estimated Gaussian distributions, allowing the Gaussian components to fit with training samples of less-frequent classes more comprehensively. Extensive experiments on commonly used datasets demonstrated that the proposed method outperforms existing methods on long-tailed image classification." @default.
• W4304080660 created "2022-10-10" @default.
• W4304080660 creator A5003019748 @default.
• W4304080660 creator A5033164669 @default.
• W4304080660 creator A5042965510 @default.
• W4304080660 creator A5060524512 @default.
• W4304080660 creator A5081226638 @default.
• W4304080660 date "2022-10-10" @default.
• W4304080660 modified "2023-10-06" @default.
• W4304080660 title "Compound Batch Normalization for Long-tailed Image Classification" @default.
• W4304080660 cites W2117539524 @default.
• W4304080660 cites W2148143831 @default.
• W4304080660 cites W2194775991 @default.
• W4304080660 cites W2440599146 @default.
• W4304080660 cites W2797977484 @default.
• W4304080660 cites W2798282242 @default.
• W4304080660 cites W2949736877 @default.
• W4304080660 cites W2963351448 @default.
• W4304080660 cites W2963691377 @default.
• W4304080660 cites W3034369739 @default.
• W4304080660 cites W3034601242 @default.
• W4304080660 cites W3035524453 @default.
• W4304080660 cites W3096121526 @default.
• W4304080660 cites W3096688134 @default.
• W4304080660 cites W3152635971 @default.
• W4304080660 cites W3166596953 @default.
• W4304080660 cites W3171007011 @default.
• W4304080660 cites W3176474016 @default.
• W4304080660 cites W3177200443 @default.
• W4304080660 cites W3177230409 @default.
• W4304080660 cites W3196349267 @default.
• W4304080660 cites W3202087803 @default.
• W4304080660 cites W3202223295 @default.
• W4304080660 cites W3202232857 @default.
• W4304080660 cites W3205497712 @default.
• W4304080660 cites W4214519401 @default.
• W4304080660 cites W4214640945 @default.
• W4304080660 cites W4214718285 @default.
• W4304080660 cites W4250482878 @default.
• W4304080660 cites W4313118515 @default.
• W4304080660 doi "https://doi.org/10.1145/3503161.3547805" @default.
• W4304080660 hasPublicationYear "2022" @default.
• W4304080660 type Work @default.
• W4304080660 citedByCount "2" @default.
• W4304080660 countsByYear W43040806602023 @default.
• W4304080660 crossrefType "proceedings-article" @default.
• W4304080660 hasAuthorship W4304080660A5003019748 @default.
• W4304080660 hasAuthorship W4304080660A5033164669 @default.
• W4304080660 hasAuthorship W4304080660A5042965510 @default.
• W4304080660 hasAuthorship W4304080660A5060524512 @default.
• W4304080660 hasAuthorship W4304080660A5081226638 @default.
• W4304080660 hasBestOaLocation W43040806602 @default.
• W4304080660 hasConcept C105795698 @default.
• W4304080660 hasConcept C114466953 @default.
• W4304080660 hasConcept C115961682 @default.
• W4304080660 hasConcept C121332964 @default.
• W4304080660 hasConcept C126838900 @default.
• W4304080660 hasConcept C134306372 @default.
• W4304080660 hasConcept C136886441 @default.
• W4304080660 hasConcept C144024400 @default.
• W4304080660 hasConcept C153180895 @default.
• W4304080660 hasConcept C154945302 @default.
• W4304080660 hasConcept C163716315 @default.
• W4304080660 hasConcept C182081679 @default.
• W4304080660 hasConcept C183115368 @default.
• W4304080660 hasConcept C185429906 @default.
• W4304080660 hasConcept C186633575 @default.
• W4304080660 hasConcept C19165224 @default.
• W4304080660 hasConcept C199360897 @default.
• W4304080660 hasConcept C202444582 @default.
• W4304080660 hasConcept C33923547 @default.
• W4304080660 hasConcept C41008148 @default.
• W4304080660 hasConcept C49781872 @default.
• W4304080660 hasConcept C52622490 @default.
• W4304080660 hasConcept C61224824 @default.
• W4304080660 hasConcept C62520636 @default.
• W4304080660 hasConcept C71924100 @default.
• W4304080660 hasConcept C75294576 @default.
• W4304080660 hasConcept C81790035 @default.
• W4304080660 hasConcept C83665646 @default.
• W4304080660 hasConcept C92757383 @default.
• W4304080660 hasConceptScore W4304080660C105795698 @default.
• W4304080660 hasConceptScore W4304080660C114466953 @default.
• W4304080660 hasConceptScore W4304080660C115961682 @default.
• W4304080660 hasConceptScore W4304080660C121332964 @default.
• W4304080660 hasConceptScore W4304080660C126838900 @default.
• W4304080660 hasConceptScore W4304080660C134306372 @default.
• W4304080660 hasConceptScore W4304080660C136886441 @default.
• W4304080660 hasConceptScore W4304080660C144024400 @default.
• W4304080660 hasConceptScore W4304080660C153180895 @default.
• W4304080660 hasConceptScore W4304080660C154945302 @default.
• W4304080660 hasConceptScore W4304080660C163716315 @default.
• W4304080660 hasConceptScore W4304080660C182081679 @default.
• W4304080660 hasConceptScore W4304080660C183115368 @default.
• W4304080660 hasConceptScore W4304080660C185429906 @default.
• W4304080660 hasConceptScore W4304080660C186633575 @default.
• W4304080660 hasConceptScore W4304080660C19165224 @default.
• W4304080660 hasConceptScore W4304080660C199360897 @default.
• W4304080660 hasConceptScore W4304080660C202444582 @default.
• W4304080660 hasConceptScore W4304080660C33923547 @default.

• next