FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3124515882> ?p ?o ?g. }

• W3124515882 abstract "Model-agnostic meta-learning (MAML) has emerged as one of the most successful meta-learning techniques in few-shot learning. It enables us to learn a meta-initialization} of model parameters (that we call meta-model) to rapidly adapt to new tasks using a small amount of labeled training data. Despite the generalization power of the meta-model, it remains elusive that how adversarial robustness can be maintained by MAML in few-shot learning. In addition to generalization, robustness is also desired for a meta-model to defend adversarial examples (attacks). Toward promoting adversarial robustness in MAML, we first study WHEN a robustness-promoting regularization should be incorporated, given the fact that MAML adopts a bi-level (fine-tuning vs. meta-update) learning procedure. We show that robustifying the meta-update stage is sufficient to make robustness adapted to the task-specific fine-tuning stage even if the latter uses a standard training protocol. We also make additional justification on the acquired robustness adaptation by peering into the interpretability of neurons' activation maps. Furthermore, we investigate HOW robust regularization can efficiently be designed in MAML. We propose a general but easily-optimized robustness-regularized meta-learning framework, which allows the use of unlabeled data augmentation, fast adversarial attack generation, and computationally-light fine-tuning. In particular, we for the first time show that the auxiliary contrastive learning task can enhance the adversarial robustness of MAML. Finally, extensive experiments are conducted to demonstrate the effectiveness of our proposed methods in robust few-shot learning." @default.
• W3124515882 created "2021-02-01" @default.
• W3124515882 creator A5002976916 @default.
• W3124515882 creator A5007306953 @default.
• W3124515882 creator A5011173717 @default.
• W3124515882 creator A5036726873 @default.
• W3124515882 creator A5040877128 @default.
• W3124515882 creator A5050344371 @default.
• W3124515882 creator A5059400886 @default.
• W3124515882 date "2021-02-20" @default.
• W3124515882 modified "2023-09-23" @default.
• W3124515882 title "On Fast Adversarial Robustness Adaptation in Model-Agnostic Meta-Learning" @default.
• W3124515882 cites W2118858186 @default.
• W3124515882 cites W2194321275 @default.
• W3124515882 cites W2601450892 @default.
• W3124515882 cites W2604763608 @default.
• W3124515882 cites W2753160622 @default.
• W3124515882 cites W2769416443 @default.
• W3124515882 cites W2795900505 @default.
• W3124515882 cites W2803392236 @default.
• W3124515882 cites W2887906331 @default.
• W3124515882 cites W2912237282 @default.
• W3124515882 cites W2913848079 @default.
• W3124515882 cites W2949362401 @default.
• W3124515882 cites W2962729158 @default.
• W3124515882 cites W2962761044 @default.
• W3124515882 cites W2962957598 @default.
• W3124515882 cites W2963207607 @default.
• W3124515882 cites W2963341924 @default.
• W3124515882 cites W2963496101 @default.
• W3124515882 cites W2963866663 @default.
• W3124515882 cites W2964206659 @default.
• W3124515882 cites W2964253222 @default.
• W3124515882 cites W2964971928 @default.
• W3124515882 cites W2970316625 @default.
• W3124515882 cites W2970317235 @default.
• W3124515882 cites W2970680991 @default.
• W3124515882 cites W2971316968 @default.
• W3124515882 cites W2974885182 @default.
• W3124515882 cites W2976501124 @default.
• W3124515882 cites W2978843034 @default.
• W3124515882 cites W2990520147 @default.
• W3124515882 cites W2996325578 @default.
• W3124515882 cites W2996564870 @default.
• W3124515882 cites W3005680577 @default.
• W3124515882 cites W3013520104 @default.
• W3124515882 cites W3034617042 @default.
• W3124515882 cites W3035106315 @default.
• W3124515882 cites W3035524453 @default.
• W3124515882 cites W3035590812 @default.
• W3124515882 cites W3041890115 @default.
• W3124515882 cites W3081534675 @default.
• W3124515882 cites W3091905774 @default.
• W3124515882 cites W3098018961 @default.
• W3124515882 cites W3098632754 @default.
• W3124515882 cites W3118062200 @default.
• W3124515882 cites W99485931 @default.
• W3124515882 doi "https://doi.org/10.48550/arxiv.2102.10454" @default.
• W3124515882 hasPublicationYear "2021" @default.
• W3124515882 type Work @default.
• W3124515882 sameAs 3124515882 @default.
• W3124515882 citedByCount "3" @default.
• W3124515882 countsByYear W31245158822021 @default.
• W3124515882 crossrefType "posted-content" @default.
• W3124515882 hasAuthorship W3124515882A5002976916 @default.
• W3124515882 hasAuthorship W3124515882A5007306953 @default.
• W3124515882 hasAuthorship W3124515882A5011173717 @default.
• W3124515882 hasAuthorship W3124515882A5036726873 @default.
• W3124515882 hasAuthorship W3124515882A5040877128 @default.
• W3124515882 hasAuthorship W3124515882A5050344371 @default.
• W3124515882 hasAuthorship W3124515882A5059400886 @default.
• W3124515882 hasBestOaLocation W31245158821 @default.
• W3124515882 hasConcept C104317684 @default.
• W3124515882 hasConcept C114466953 @default.
• W3124515882 hasConcept C119857082 @default.
• W3124515882 hasConcept C154945302 @default.
• W3124515882 hasConcept C185592680 @default.
• W3124515882 hasConcept C199360897 @default.
• W3124515882 hasConcept C2776135515 @default.
• W3124515882 hasConcept C2781067378 @default.
• W3124515882 hasConcept C37736160 @default.
• W3124515882 hasConcept C41008148 @default.
• W3124515882 hasConcept C55493867 @default.
• W3124515882 hasConcept C63479239 @default.
• W3124515882 hasConceptScore W3124515882C104317684 @default.
• W3124515882 hasConceptScore W3124515882C114466953 @default.
• W3124515882 hasConceptScore W3124515882C119857082 @default.
• W3124515882 hasConceptScore W3124515882C154945302 @default.
• W3124515882 hasConceptScore W3124515882C185592680 @default.
• W3124515882 hasConceptScore W3124515882C199360897 @default.
• W3124515882 hasConceptScore W3124515882C2776135515 @default.
• W3124515882 hasConceptScore W3124515882C2781067378 @default.
• W3124515882 hasConceptScore W3124515882C37736160 @default.
• W3124515882 hasConceptScore W3124515882C41008148 @default.
• W3124515882 hasConceptScore W3124515882C55493867 @default.
• W3124515882 hasConceptScore W3124515882C63479239 @default.
• W3124515882 hasLocation W31245158821 @default.
• W3124515882 hasOpenAccess W3124515882 @default.
• W3124515882 hasPrimaryLocation W31245158821 @default.
• W3124515882 hasRelatedWork W2947042731 @default.

• next