Matches in SemOpenAlex for { <https://semopenalex.org/work/W4312694987> ?p ?o ?g. }
Showing items 1 to 81 of
81
with 100 items per page.
- W4312694987 endingPage "529" @default.
- W4312694987 startingPage "513" @default.
- W4312694987 abstract "Pre-trained representation is one of the key elements in the success of modern deep learning. However, existing works on continual learning methods have mostly focused on learning models incrementally from scratch. In this paper, we explore an alternative framework to incremental learning where we continually fine-tune the model from a pre-trained representation. Our method takes advantage of linearization technique of a pre-trained neural network for simple and effective continual learning. We show that this allows us to design a linear model where quadratic parameter regularization method is placed as the optimal continual learning policy, and at the same time enjoying the high performance of neural networks. We also show that the proposed algorithm enables parameter regularization methods to be applied to class-incremental problems. Additionally, we provide a theoretical reason why the existing parameter-space regularization algorithms such as EWC underperform on neural networks trained with cross-entropy loss. We show that the proposed method can prevent forgetting while achieving high continual fine-tuning performance on image classification tasks. To show that our method can be applied to general continual learning settings, we evaluate our method in data-incremental, task-incremental, and class-incremental learning problems." @default.
- W4312694987 created "2023-01-05" @default.
- W4312694987 creator A5033951605 @default.
- W4312694987 creator A5060762204 @default.
- W4312694987 creator A5066874418 @default.
- W4312694987 creator A5083790649 @default.
- W4312694987 date "2022-01-01" @default.
- W4312694987 modified "2023-10-18" @default.
- W4312694987 title "DLCFT: Deep Linear Continual Fine-Tuning for General Incremental Learning" @default.
- W4312694987 cites W2117539524 @default.
- W4312694987 cites W2152161678 @default.
- W4312694987 cites W2194775991 @default.
- W4312694987 cites W2473930607 @default.
- W4312694987 cites W2554616628 @default.
- W4312694987 cites W2560647685 @default.
- W4312694987 cites W2962858109 @default.
- W4312694987 cites W2963072899 @default.
- W4312694987 cites W2963588172 @default.
- W4312694987 cites W2963788399 @default.
- W4312694987 cites W2964189064 @default.
- W4312694987 cites W2972313371 @default.
- W4312694987 cites W3030364939 @default.
- W4312694987 cites W3034736335 @default.
- W4312694987 cites W3097784654 @default.
- W4312694987 cites W3175141461 @default.
- W4312694987 cites W3176509493 @default.
- W4312694987 doi "https://doi.org/10.1007/978-3-031-19827-4_30" @default.
- W4312694987 hasPublicationYear "2022" @default.
- W4312694987 type Work @default.
- W4312694987 citedByCount "0" @default.
- W4312694987 crossrefType "book-chapter" @default.
- W4312694987 hasAuthorship W4312694987A5033951605 @default.
- W4312694987 hasAuthorship W4312694987A5060762204 @default.
- W4312694987 hasAuthorship W4312694987A5066874418 @default.
- W4312694987 hasAuthorship W4312694987A5083790649 @default.
- W4312694987 hasBestOaLocation W43126949872 @default.
- W4312694987 hasConcept C108583219 @default.
- W4312694987 hasConcept C11210021 @default.
- W4312694987 hasConcept C119857082 @default.
- W4312694987 hasConcept C121332964 @default.
- W4312694987 hasConcept C138885662 @default.
- W4312694987 hasConcept C154945302 @default.
- W4312694987 hasConcept C158622935 @default.
- W4312694987 hasConcept C2776135515 @default.
- W4312694987 hasConcept C41008148 @default.
- W4312694987 hasConcept C41895202 @default.
- W4312694987 hasConcept C50644808 @default.
- W4312694987 hasConcept C62520636 @default.
- W4312694987 hasConcept C7149132 @default.
- W4312694987 hasConceptScore W4312694987C108583219 @default.
- W4312694987 hasConceptScore W4312694987C11210021 @default.
- W4312694987 hasConceptScore W4312694987C119857082 @default.
- W4312694987 hasConceptScore W4312694987C121332964 @default.
- W4312694987 hasConceptScore W4312694987C138885662 @default.
- W4312694987 hasConceptScore W4312694987C154945302 @default.
- W4312694987 hasConceptScore W4312694987C158622935 @default.
- W4312694987 hasConceptScore W4312694987C2776135515 @default.
- W4312694987 hasConceptScore W4312694987C41008148 @default.
- W4312694987 hasConceptScore W4312694987C41895202 @default.
- W4312694987 hasConceptScore W4312694987C50644808 @default.
- W4312694987 hasConceptScore W4312694987C62520636 @default.
- W4312694987 hasConceptScore W4312694987C7149132 @default.
- W4312694987 hasLocation W43126949871 @default.
- W4312694987 hasLocation W43126949872 @default.
- W4312694987 hasOpenAccess W4312694987 @default.
- W4312694987 hasPrimaryLocation W43126949871 @default.
- W4312694987 hasRelatedWork W2922457425 @default.
- W4312694987 hasRelatedWork W3014300295 @default.
- W4312694987 hasRelatedWork W3164822677 @default.
- W4312694987 hasRelatedWork W4223943233 @default.
- W4312694987 hasRelatedWork W4225161397 @default.
- W4312694987 hasRelatedWork W4250304930 @default.
- W4312694987 hasRelatedWork W4309045103 @default.
- W4312694987 hasRelatedWork W4312200629 @default.
- W4312694987 hasRelatedWork W4360585206 @default.
- W4312694987 hasRelatedWork W4364306694 @default.
- W4312694987 isParatext "false" @default.
- W4312694987 isRetracted "false" @default.
- W4312694987 workType "book-chapter" @default.