FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3125116114> ?p ?o ?g. }

• W3125116114 endingPage "51" @default.
• W3125116114 startingPage "28" @default.
• W3125116114 abstract "Online continual learning for image classification studies the problem of learning to classify images from an online stream of data and tasks, where tasks may include new classes (class incremental) or data nonstationarity (domain incremental). One of the key challenges of continual learning is to avoid catastrophic forgetting (CF), i.e., forgetting old tasks in the presence of more recent tasks. Over the past few years, a large range of methods and tricks have been introduced to address the continual learning problem, but many have not been fairly and systematically compared under a variety of realistic and practical settings. To better understand the relative advantages of various approaches and the settings where they work best, this survey aims to (1) compare state-of-the-art methods such as Maximally Interfered Retrieval (MIR), iCARL, and GDumb (a very strong baseline) and determine which works best at different memory and data settings as well as better understand the key source of CF; (2) determine if the best online class incremental methods are also competitive in the domain incremental setting; and (3) evaluate the performance of 7 simple but effective tricks such as the ”review” trick and the nearest class mean (NCM) classifier to assess their relative impact. Regarding (1), we observe that iCaRL remains competitive when the memory buffer is small; GDumb outperforms many recently proposed methods in medium-size datasets and MIR performs the best in larger-scale datasets. For (2), we note that GDumb performs quite poorly while MIR – already competitive for (1) – is also strongly competitive in this very different (but important) continual learning setting. Overall, this allows us to conclude that MIR is overall a strong and versatile online continual learning method across a wide variety of settings. Finally for (3), we find that all tricks are beneficial, and when augmented with the “review” trick and NCM classifier, MIR produces performance levels that bring online continual learning much closer to its ultimate goal of matching offline training. Our codes are available at https://github.com/RaptorMai/online-continual-learning." @default.
• W3125116114 created "2021-02-01" @default.
• W3125116114 creator A5008383067 @default.
• W3125116114 creator A5027791813 @default.
• W3125116114 creator A5028174137 @default.
• W3125116114 creator A5054396441 @default.
• W3125116114 creator A5062110178 @default.
• W3125116114 creator A5066572795 @default.
• W3125116114 date "2022-01-01" @default.
• W3125116114 modified "2023-10-03" @default.
• W3125116114 title "Online continual learning in image classification: An empirical survey" @default.
• W3125116114 cites W1528242606 @default.
• W3125116114 cites W1682403713 @default.
• W3125116114 cites W2015563892 @default.
• W3125116114 cites W2037979274 @default.
• W3125116114 cites W2105628133 @default.
• W3125116114 cites W2112796928 @default.
• W3125116114 cites W2119885577 @default.
• W3125116114 cites W2144219012 @default.
• W3125116114 cites W2560647685 @default.
• W3125116114 cites W2565989828 @default.
• W3125116114 cites W2788388592 @default.
• W3125116114 cites W2798869704 @default.
• W3125116114 cites W2962884963 @default.
• W3125116114 cites W2963876099 @default.
• W3125116114 cites W2964631355 @default.
• W3125116114 cites W2972566450 @default.
• W3125116114 cites W2974859516 @default.
• W3125116114 cites W2977932430 @default.
• W3125116114 cites W2983826605 @default.
• W3125116114 cites W2996514457 @default.
• W3125116114 cites W2996772410 @default.
• W3125116114 cites W3012711517 @default.
• W3125116114 cites W3021931813 @default.
• W3125116114 cites W3031989616 @default.
• W3125116114 cites W3035086011 @default.
• W3125116114 cites W3094953068 @default.
• W3125116114 cites W3110148193 @default.
• W3125116114 cites W3130320103 @default.
• W3125116114 cites W3163842339 @default.
• W3125116114 cites W3179436402 @default.
• W3125116114 cites W3186185618 @default.
• W3125116114 doi "https://doi.org/10.1016/j.neucom.2021.10.021" @default.
• W3125116114 hasPublicationYear "2022" @default.
• W3125116114 type Work @default.
• W3125116114 sameAs 3125116114 @default.
• W3125116114 citedByCount "88" @default.
• W3125116114 countsByYear W31251161142020 @default.
• W3125116114 countsByYear W31251161142021 @default.
• W3125116114 countsByYear W31251161142022 @default.
• W3125116114 countsByYear W31251161142023 @default.
• W3125116114 crossrefType "journal-article" @default.
• W3125116114 hasAuthorship W3125116114A5008383067 @default.
• W3125116114 hasAuthorship W3125116114A5027791813 @default.
• W3125116114 hasAuthorship W3125116114A5028174137 @default.
• W3125116114 hasAuthorship W3125116114A5054396441 @default.
• W3125116114 hasAuthorship W3125116114A5062110178 @default.
• W3125116114 hasAuthorship W3125116114A5066572795 @default.
• W3125116114 hasBestOaLocation W31251161142 @default.
• W3125116114 hasConcept C111368507 @default.
• W3125116114 hasConcept C119857082 @default.
• W3125116114 hasConcept C12725497 @default.
• W3125116114 hasConcept C127313418 @default.
• W3125116114 hasConcept C138885662 @default.
• W3125116114 hasConcept C154945302 @default.
• W3125116114 hasConcept C26517878 @default.
• W3125116114 hasConcept C2777212361 @default.
• W3125116114 hasConcept C2780735816 @default.
• W3125116114 hasConcept C38652104 @default.
• W3125116114 hasConcept C41008148 @default.
• W3125116114 hasConcept C41895202 @default.
• W3125116114 hasConcept C7149132 @default.
• W3125116114 hasConcept C95623464 @default.
• W3125116114 hasConceptScore W3125116114C111368507 @default.
• W3125116114 hasConceptScore W3125116114C119857082 @default.
• W3125116114 hasConceptScore W3125116114C12725497 @default.
• W3125116114 hasConceptScore W3125116114C127313418 @default.
• W3125116114 hasConceptScore W3125116114C138885662 @default.
• W3125116114 hasConceptScore W3125116114C154945302 @default.
• W3125116114 hasConceptScore W3125116114C26517878 @default.
• W3125116114 hasConceptScore W3125116114C2777212361 @default.
• W3125116114 hasConceptScore W3125116114C2780735816 @default.
• W3125116114 hasConceptScore W3125116114C38652104 @default.
• W3125116114 hasConceptScore W3125116114C41008148 @default.
• W3125116114 hasConceptScore W3125116114C41895202 @default.
• W3125116114 hasConceptScore W3125116114C7149132 @default.
• W3125116114 hasConceptScore W3125116114C95623464 @default.
• W3125116114 hasLocation W31251161141 @default.
• W3125116114 hasLocation W31251161142 @default.
• W3125116114 hasOpenAccess W3125116114 @default.
• W3125116114 hasPrimaryLocation W31251161141 @default.
• W3125116114 hasRelatedWork W1570833164 @default.
• W3125116114 hasRelatedWork W2961085424 @default.
• W3125116114 hasRelatedWork W3034933965 @default.
• W3125116114 hasRelatedWork W3088215962 @default.
• W3125116114 hasRelatedWork W3126776133 @default.
• W3125116114 hasRelatedWork W3127669944 @default.
• W3125116114 hasRelatedWork W3173986289 @default.

• next