Matches in SemOpenAlex for { <https://semopenalex.org/work/W3014423797> ?p ?o ?g. }
- W3014423797 abstract "Real-world visual recognition requires handling the extreme sample imbalance in large-scale long-tailed data. We propose a divide&conquer strategy for the challenging LVIS task: divide the whole data into balanced parts and then apply incremental learning to conquer each one. This derives a novel learning paradigm: class-incremental few-shot learning, which is especially effective for the challenge evolving over time: 1) the class imbalance among the old-class knowledge review and 2) the few-shot data in new-class learning. We call our approach Learning to Segment the Tail (LST). In particular, we design an instance-level balanced replay scheme, which is a memory-efficient approximation to balance the instance-level samples from the old-class images. We also propose to use a meta-module for new-class learning, where the module parameters are shared across incremental phases, gaining the learning-to-learn knowledge incrementally, from the data-rich head to the data-poor tail. We empirically show that: at the expense of a little sacrifice of head-class forgetting, we can gain a significant 8.3% AP improvement for the tail classes with less than 10 instances, achieving an overall 2.0% AP boost for the whole 1,230 classes." @default.
- W3014423797 created "2020-04-10" @default.
- W3014423797 creator A5001984245 @default.
- W3014423797 creator A5003190764 @default.
- W3014423797 creator A5039896971 @default.
- W3014423797 creator A5042324027 @default.
- W3014423797 creator A5044120384 @default.
- W3014423797 creator A5083753074 @default.
- W3014423797 date "2020-04-02" @default.
- W3014423797 modified "2023-09-24" @default.
- W3014423797 title "Learning to Segment the Tail" @default.
- W3014423797 cites W1509754072 @default.
- W3014423797 cites W1536680647 @default.
- W3014423797 cites W1682403713 @default.
- W3014423797 cites W1821462560 @default.
- W3014423797 cites W1861492603 @default.
- W3014423797 cites W1968625547 @default.
- W3014423797 cites W2017814585 @default.
- W3014423797 cites W2053782355 @default.
- W3014423797 cites W2104933073 @default.
- W3014423797 cites W2107742354 @default.
- W3014423797 cites W2110764733 @default.
- W3014423797 cites W2118978333 @default.
- W3014423797 cites W2187089797 @default.
- W3014423797 cites W2194775991 @default.
- W3014423797 cites W2339172597 @default.
- W3014423797 cites W2440599146 @default.
- W3014423797 cites W2549139847 @default.
- W3014423797 cites W2555182955 @default.
- W3014423797 cites W2613718673 @default.
- W3014423797 cites W2753300133 @default.
- W3014423797 cites W2775447965 @default.
- W3014423797 cites W2788313038 @default.
- W3014423797 cites W2796346823 @default.
- W3014423797 cites W2798836702 @default.
- W3014423797 cites W2884282566 @default.
- W3014423797 cites W2948672349 @default.
- W3014423797 cites W2948734064 @default.
- W3014423797 cites W2948879658 @default.
- W3014423797 cites W2962966271 @default.
- W3014423797 cites W2963150697 @default.
- W3014423797 cites W2963341924 @default.
- W3014423797 cites W2963691377 @default.
- W3014423797 cites W2963849369 @default.
- W3014423797 cites W2963857746 @default.
- W3014423797 cites W2963921921 @default.
- W3014423797 cites W2963943197 @default.
- W3014423797 cites W2964189064 @default.
- W3014423797 cites W2964236837 @default.
- W3014423797 cites W2966993394 @default.
- W3014423797 cites W2970689606 @default.
- W3014423797 cites W2972231924 @default.
- W3014423797 cites W2983156430 @default.
- W3014423797 cites W2989701968 @default.
- W3014423797 cites W2997021253 @default.
- W3014423797 cites W3012424925 @default.
- W3014423797 cites W3021931813 @default.
- W3014423797 cites W3100156920 @default.
- W3014423797 cites W72187777 @default.
- W3014423797 hasPublicationYear "2020" @default.
- W3014423797 type Work @default.
- W3014423797 sameAs 3014423797 @default.
- W3014423797 citedByCount "2" @default.
- W3014423797 countsByYear W30144237972020 @default.
- W3014423797 countsByYear W30144237972021 @default.
- W3014423797 crossrefType "posted-content" @default.
- W3014423797 hasAuthorship W3014423797A5001984245 @default.
- W3014423797 hasAuthorship W3014423797A5003190764 @default.
- W3014423797 hasAuthorship W3014423797A5039896971 @default.
- W3014423797 hasAuthorship W3014423797A5042324027 @default.
- W3014423797 hasAuthorship W3014423797A5044120384 @default.
- W3014423797 hasAuthorship W3014423797A5083753074 @default.
- W3014423797 hasConcept C11413529 @default.
- W3014423797 hasConcept C119857082 @default.
- W3014423797 hasConcept C138885662 @default.
- W3014423797 hasConcept C154945302 @default.
- W3014423797 hasConcept C162324750 @default.
- W3014423797 hasConcept C187736073 @default.
- W3014423797 hasConcept C2777212361 @default.
- W3014423797 hasConcept C2780451532 @default.
- W3014423797 hasConcept C2780735816 @default.
- W3014423797 hasConcept C41008148 @default.
- W3014423797 hasConcept C41895202 @default.
- W3014423797 hasConcept C7149132 @default.
- W3014423797 hasConcept C71559656 @default.
- W3014423797 hasConceptScore W3014423797C11413529 @default.
- W3014423797 hasConceptScore W3014423797C119857082 @default.
- W3014423797 hasConceptScore W3014423797C138885662 @default.
- W3014423797 hasConceptScore W3014423797C154945302 @default.
- W3014423797 hasConceptScore W3014423797C162324750 @default.
- W3014423797 hasConceptScore W3014423797C187736073 @default.
- W3014423797 hasConceptScore W3014423797C2777212361 @default.
- W3014423797 hasConceptScore W3014423797C2780451532 @default.
- W3014423797 hasConceptScore W3014423797C2780735816 @default.
- W3014423797 hasConceptScore W3014423797C41008148 @default.
- W3014423797 hasConceptScore W3014423797C41895202 @default.
- W3014423797 hasConceptScore W3014423797C7149132 @default.
- W3014423797 hasConceptScore W3014423797C71559656 @default.
- W3014423797 hasLocation W30144237971 @default.
- W3014423797 hasOpenAccess W3014423797 @default.