Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386918742> ?p ?o ?g. }
Showing items 1 to 72 of
72
with 100 items per page.
- W4386918742 endingPage "16" @default.
- W4386918742 startingPage "1" @default.
- W4386918742 abstract "Continual learning (CL) aims to learn a non-stationary data distribution and not forget previous knowledge. The effectiveness of existing approaches that rely on memory replay can decrease over time as the model tends to overfit the stored examples. As a result, the model's ability to generalize well is significantly constrained. Additionally, these methods often overlook the inherent uncertainty in the memory data distribution, which differs significantly from the distribution of all previous data examples. To overcome these issues, we propose a principled memory evolution framework that dynamically adjusts the memory data distribution. This evolution is achieved by employing distributionally robust optimization (DRO) to make the memory buffer increasingly difficult to memorize. We consider two types of constraints in DRO: f-divergence and Wasserstein ball constraints. For f-divergence constraint, we derive a family of methods to evolve the memory buffer data in the continuous probability measure space with Wasserstein gradient flow (WGF). For Wasserstein ball constraint, we directly solve it in the Euclidean space. Extensive experiments on existing benchmarks demonstrate the effectiveness of the proposed methods for alleviating forgetting. As a by-product of the proposed framework, our method is more robust to adversarial examples than compared CL methods." @default.
- W4386918742 created "2023-09-22" @default.
- W4386918742 creator A5000950850 @default.
- W4386918742 creator A5024023588 @default.
- W4386918742 creator A5033311476 @default.
- W4386918742 creator A5039291107 @default.
- W4386918742 creator A5042802004 @default.
- W4386918742 creator A5044277083 @default.
- W4386918742 creator A5071037763 @default.
- W4386918742 date "2023-01-01" @default.
- W4386918742 modified "2023-09-29" @default.
- W4386918742 title "Distributionally Robust Memory Evolution With Generalized Divergence for Continual Learning" @default.
- W4386918742 doi "https://doi.org/10.1109/tpami.2023.3317874" @default.
- W4386918742 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37738203" @default.
- W4386918742 hasPublicationYear "2023" @default.
- W4386918742 type Work @default.
- W4386918742 citedByCount "0" @default.
- W4386918742 crossrefType "journal-article" @default.
- W4386918742 hasAuthorship W4386918742A5000950850 @default.
- W4386918742 hasAuthorship W4386918742A5024023588 @default.
- W4386918742 hasAuthorship W4386918742A5033311476 @default.
- W4386918742 hasAuthorship W4386918742A5039291107 @default.
- W4386918742 hasAuthorship W4386918742A5042802004 @default.
- W4386918742 hasAuthorship W4386918742A5044277083 @default.
- W4386918742 hasAuthorship W4386918742A5071037763 @default.
- W4386918742 hasConcept C105795698 @default.
- W4386918742 hasConcept C126255220 @default.
- W4386918742 hasConcept C138885662 @default.
- W4386918742 hasConcept C149441793 @default.
- W4386918742 hasConcept C154945302 @default.
- W4386918742 hasConcept C207390915 @default.
- W4386918742 hasConcept C22019652 @default.
- W4386918742 hasConcept C2524010 @default.
- W4386918742 hasConcept C2776036281 @default.
- W4386918742 hasConcept C33923547 @default.
- W4386918742 hasConcept C41008148 @default.
- W4386918742 hasConcept C41895202 @default.
- W4386918742 hasConcept C50644808 @default.
- W4386918742 hasConcept C7149132 @default.
- W4386918742 hasConceptScore W4386918742C105795698 @default.
- W4386918742 hasConceptScore W4386918742C126255220 @default.
- W4386918742 hasConceptScore W4386918742C138885662 @default.
- W4386918742 hasConceptScore W4386918742C149441793 @default.
- W4386918742 hasConceptScore W4386918742C154945302 @default.
- W4386918742 hasConceptScore W4386918742C207390915 @default.
- W4386918742 hasConceptScore W4386918742C22019652 @default.
- W4386918742 hasConceptScore W4386918742C2524010 @default.
- W4386918742 hasConceptScore W4386918742C2776036281 @default.
- W4386918742 hasConceptScore W4386918742C33923547 @default.
- W4386918742 hasConceptScore W4386918742C41008148 @default.
- W4386918742 hasConceptScore W4386918742C41895202 @default.
- W4386918742 hasConceptScore W4386918742C50644808 @default.
- W4386918742 hasConceptScore W4386918742C7149132 @default.
- W4386918742 hasLocation W43869187421 @default.
- W4386918742 hasLocation W43869187422 @default.
- W4386918742 hasOpenAccess W4386918742 @default.
- W4386918742 hasPrimaryLocation W43869187421 @default.
- W4386918742 hasRelatedWork W2792213864 @default.
- W4386918742 hasRelatedWork W2989932438 @default.
- W4386918742 hasRelatedWork W3099765033 @default.
- W4386918742 hasRelatedWork W3170349722 @default.
- W4386918742 hasRelatedWork W3172676853 @default.
- W4386918742 hasRelatedWork W3190073114 @default.
- W4386918742 hasRelatedWork W4283827158 @default.
- W4386918742 hasRelatedWork W4287120308 @default.
- W4386918742 hasRelatedWork W4376312092 @default.
- W4386918742 hasRelatedWork W4385571156 @default.
- W4386918742 isParatext "false" @default.
- W4386918742 isRetracted "false" @default.
- W4386918742 workType "article" @default.