Matches in SemOpenAlex for { <https://semopenalex.org/work/W3206454299> ?p ?o ?g. }
- W3206454299 abstract "Continual learning (CL) aims to learn a sequence of tasks without forgetting the previously acquired knowledge. However, recent advances in continual learning are restricted to supervised continual learning (SCL) scenarios. Consequently, they are not scalable to real-world applications where the data distribution is often biased and unannotated. In this work, we focus on unsupervised continual learning (UCL), where we learn the feature representations on an unlabelled sequence of tasks and show that reliance on annotated data is not necessary for continual learning. We conduct a systematic study analyzing the learned feature representations and show that unsupervised visual representations are surprisingly more robust to catastrophic forgetting, consistently achieve better performance, and generalize better to out-of-distribution tasks than SCL. Furthermore, we find that UCL achieves a smoother loss landscape through qualitative analysis of the learned representations and learns meaningful feature representations. Additionally, we propose Lifelong Unsupervised Mixup (LUMP), a simple yet effective technique that leverages the interpolation between the current task and previous tasks' instances to alleviate catastrophic forgetting for unsupervised representations." @default.
- W3206454299 created "2021-10-25" @default.
- W3206454299 creator A5014456335 @default.
- W3206454299 creator A5027782298 @default.
- W3206454299 creator A5039876507 @default.
- W3206454299 creator A5054456399 @default.
- W3206454299 creator A5070302452 @default.
- W3206454299 date "2021-10-13" @default.
- W3206454299 modified "2023-10-06" @default.
- W3206454299 title "Rethinking the Representational Continuity: Towards Unsupervised Continual Learning." @default.
- W3206454299 cites W1682403713 @default.
- W3206454299 cites W1942758450 @default.
- W3206454299 cites W2108598243 @default.
- W3206454299 cites W2113839990 @default.
- W3206454299 cites W2171590421 @default.
- W3206454299 cites W2194321275 @default.
- W3206454299 cites W2194775991 @default.
- W3206454299 cites W2335728318 @default.
- W3206454299 cites W2342840547 @default.
- W3206454299 cites W2604763608 @default.
- W3206454299 cites W2737492962 @default.
- W3206454299 cites W2750384547 @default.
- W3206454299 cites W2798991696 @default.
- W3206454299 cites W2896084542 @default.
- W3206454299 cites W2902625698 @default.
- W3206454299 cites W2926477959 @default.
- W3206454299 cites W2942810103 @default.
- W3206454299 cites W2962933129 @default.
- W3206454299 cites W2963399829 @default.
- W3206454299 cites W2963438784 @default.
- W3206454299 cites W2963540014 @default.
- W3206454299 cites W2963850662 @default.
- W3206454299 cites W2964189064 @default.
- W3206454299 cites W2970066656 @default.
- W3206454299 cites W2970502342 @default.
- W3206454299 cites W2970586779 @default.
- W3206454299 cites W2970748152 @default.
- W3206454299 cites W2982580298 @default.
- W3206454299 cites W2996486651 @default.
- W3206454299 cites W3009561768 @default.
- W3206454299 cites W3010801434 @default.
- W3206454299 cites W3021931813 @default.
- W3206454299 cites W3034549794 @default.
- W3206454299 cites W3034978746 @default.
- W3206454299 cites W3035524453 @default.
- W3206454299 cites W3095121901 @default.
- W3206454299 cites W3098033339 @default.
- W3206454299 cites W3099495704 @default.
- W3206454299 cites W3100859887 @default.
- W3206454299 cites W3101821705 @default.
- W3206454299 cites W3103936143 @default.
- W3206454299 cites W3106522756 @default.
- W3206454299 cites W3118608800 @default.
- W3206454299 cites W3125645205 @default.
- W3206454299 cites W3171007011 @default.
- W3206454299 cites W81409850 @default.
- W3206454299 hasPublicationYear "2021" @default.
- W3206454299 type Work @default.
- W3206454299 sameAs 3206454299 @default.
- W3206454299 citedByCount "0" @default.
- W3206454299 crossrefType "posted-content" @default.
- W3206454299 hasAuthorship W3206454299A5014456335 @default.
- W3206454299 hasAuthorship W3206454299A5027782298 @default.
- W3206454299 hasAuthorship W3206454299A5039876507 @default.
- W3206454299 hasAuthorship W3206454299A5054456399 @default.
- W3206454299 hasAuthorship W3206454299A5070302452 @default.
- W3206454299 hasConcept C119857082 @default.
- W3206454299 hasConcept C120665830 @default.
- W3206454299 hasConcept C121332964 @default.
- W3206454299 hasConcept C138885662 @default.
- W3206454299 hasConcept C154945302 @default.
- W3206454299 hasConcept C15744967 @default.
- W3206454299 hasConcept C162324750 @default.
- W3206454299 hasConcept C180747234 @default.
- W3206454299 hasConcept C187736073 @default.
- W3206454299 hasConcept C192209626 @default.
- W3206454299 hasConcept C2776401178 @default.
- W3206454299 hasConcept C2778112365 @default.
- W3206454299 hasConcept C2780451532 @default.
- W3206454299 hasConcept C41008148 @default.
- W3206454299 hasConcept C41895202 @default.
- W3206454299 hasConcept C48044578 @default.
- W3206454299 hasConcept C54355233 @default.
- W3206454299 hasConcept C59404180 @default.
- W3206454299 hasConcept C7149132 @default.
- W3206454299 hasConcept C77088390 @default.
- W3206454299 hasConcept C8038995 @default.
- W3206454299 hasConcept C86803240 @default.
- W3206454299 hasConceptScore W3206454299C119857082 @default.
- W3206454299 hasConceptScore W3206454299C120665830 @default.
- W3206454299 hasConceptScore W3206454299C121332964 @default.
- W3206454299 hasConceptScore W3206454299C138885662 @default.
- W3206454299 hasConceptScore W3206454299C154945302 @default.
- W3206454299 hasConceptScore W3206454299C15744967 @default.
- W3206454299 hasConceptScore W3206454299C162324750 @default.
- W3206454299 hasConceptScore W3206454299C180747234 @default.
- W3206454299 hasConceptScore W3206454299C187736073 @default.
- W3206454299 hasConceptScore W3206454299C192209626 @default.
- W3206454299 hasConceptScore W3206454299C2776401178 @default.
- W3206454299 hasConceptScore W3206454299C2778112365 @default.