Matches in SemOpenAlex for { <https://semopenalex.org/work/W4313180564> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W4313180564 endingPage "442" @default.
- W4313180564 startingPage "426" @default.
- W4313180564 abstract "Deep long-tailed learning aims to train useful deep networks on practical, real-world imbalanced distributions, wherein most labels of the tail classes are associated with a few samples. There has been a large body of work to train discriminative models for visual recognition on long-tailed distribution. In contrast, we aim to train conditional Generative Adversarial Networks, a class of image generation models on long-tailed distributions. We find that similar to recognition, state-of-the-art methods for image generation also suffer from performance degradation on tail classes. The performance degradation is mainly due to class-specific mode collapse for tail classes, which we observe to be correlated with the spectral explosion of the conditioning parameter matrix. We propose a novel group Spectral Regularizer (gSR) that prevents the spectral explosion alleviating mode collapse, which results in diverse and plausible image generation even for tail classes. We find that gSR effectively combines with existing augmentation and regularization techniques, leading to state-of-the-art image generation performance on long-tailed data. Extensive experiments demonstrate the efficacy of our regularizer on long-tailed datasets with different degrees of imbalance. Project Page: https://sites.google.com/view/gsr-eccv22 ." @default.
- W4313180564 created "2023-01-06" @default.
- W4313180564 creator A5010755999 @default.
- W4313180564 creator A5061892438 @default.
- W4313180564 creator A5062096014 @default.
- W4313180564 creator A5081429561 @default.
- W4313180564 creator A5085870393 @default.
- W4313180564 date "2022-01-01" @default.
- W4313180564 modified "2023-10-14" @default.
- W4313180564 title "Improving GANs for Long-Tailed Data Through Group Spectral Regularization" @default.
- W4313180564 cites W2108598243 @default.
- W4313180564 cites W2130481596 @default.
- W4313180564 cites W2593414223 @default.
- W4313180564 cites W2963306805 @default.
- W4313180564 cites W2963470893 @default.
- W4313180564 cites W2963691377 @default.
- W4313180564 cites W2964334477 @default.
- W4313180564 cites W2981515171 @default.
- W4313180564 cites W2990251202 @default.
- W4313180564 cites W3034601242 @default.
- W4313180564 cites W3120254195 @default.
- W4313180564 cites W3126434301 @default.
- W4313180564 cites W3171615848 @default.
- W4313180564 cites W3175491752 @default.
- W4313180564 cites W3175636332 @default.
- W4313180564 cites W4214669978 @default.
- W4313180564 cites W4250482878 @default.
- W4313180564 doi "https://doi.org/10.1007/978-3-031-19784-0_25" @default.
- W4313180564 hasPublicationYear "2022" @default.
- W4313180564 type Work @default.
- W4313180564 citedByCount "1" @default.
- W4313180564 countsByYear W43131805642023 @default.
- W4313180564 crossrefType "book-chapter" @default.
- W4313180564 hasAuthorship W4313180564A5010755999 @default.
- W4313180564 hasAuthorship W4313180564A5061892438 @default.
- W4313180564 hasAuthorship W4313180564A5062096014 @default.
- W4313180564 hasAuthorship W4313180564A5081429561 @default.
- W4313180564 hasAuthorship W4313180564A5085870393 @default.
- W4313180564 hasBestOaLocation W43131805642 @default.
- W4313180564 hasConcept C108583219 @default.
- W4313180564 hasConcept C115961682 @default.
- W4313180564 hasConcept C119857082 @default.
- W4313180564 hasConcept C153180895 @default.
- W4313180564 hasConcept C154945302 @default.
- W4313180564 hasConcept C2776135515 @default.
- W4313180564 hasConcept C39890363 @default.
- W4313180564 hasConcept C41008148 @default.
- W4313180564 hasConcept C97931131 @default.
- W4313180564 hasConceptScore W4313180564C108583219 @default.
- W4313180564 hasConceptScore W4313180564C115961682 @default.
- W4313180564 hasConceptScore W4313180564C119857082 @default.
- W4313180564 hasConceptScore W4313180564C153180895 @default.
- W4313180564 hasConceptScore W4313180564C154945302 @default.
- W4313180564 hasConceptScore W4313180564C2776135515 @default.
- W4313180564 hasConceptScore W4313180564C39890363 @default.
- W4313180564 hasConceptScore W4313180564C41008148 @default.
- W4313180564 hasConceptScore W4313180564C97931131 @default.
- W4313180564 hasLocation W43131805641 @default.
- W4313180564 hasLocation W43131805642 @default.
- W4313180564 hasLocation W43131805643 @default.
- W4313180564 hasOpenAccess W4313180564 @default.
- W4313180564 hasPrimaryLocation W43131805641 @default.
- W4313180564 hasRelatedWork W1652783584 @default.
- W4313180564 hasRelatedWork W1990254706 @default.
- W4313180564 hasRelatedWork W2024160000 @default.
- W4313180564 hasRelatedWork W2353457699 @default.
- W4313180564 hasRelatedWork W2404514746 @default.
- W4313180564 hasRelatedWork W2729514902 @default.
- W4313180564 hasRelatedWork W2743258233 @default.
- W4313180564 hasRelatedWork W2773500201 @default.
- W4313180564 hasRelatedWork W2998168123 @default.
- W4313180564 hasRelatedWork W4287995534 @default.
- W4313180564 isParatext "false" @default.
- W4313180564 isRetracted "false" @default.
- W4313180564 workType "book-chapter" @default.