Learning to Balance: Bayesian Meta-Learning for Imbalanced and Out-of-distribution Tasks

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 90
  • Download : 0
DC FieldValueLanguage
dc.contributor.author이해범ko
dc.contributor.author이하연ko
dc.contributor.authorNa, Donghyunko
dc.contributor.authorKim, Saehoonko
dc.contributor.authorPark, Minseopko
dc.contributor.authorYang, Eunhoko
dc.contributor.authorHwang, Sung Juko
dc.date.accessioned2020-12-30T04:50:29Z-
dc.date.available2020-12-30T04:50:29Z-
dc.date.created2020-12-03-
dc.date.created2020-12-03-
dc.date.issued2020-04-27-
dc.identifier.citationEighth International Conference on Learning Representations, ICLR 2020-
dc.identifier.urihttp://hdl.handle.net/10203/279338-
dc.description.abstractWhile tasks could come with varying the number of instances and classes in realistic settings, the existing meta-learning approaches for few-shot classification assume that the number of instances per task and class is fixed. Due to such restriction, they learn to equally utilize the meta-knowledge across all the tasks, even when the number of instances per task and class largely varies. Moreover, they do not consider distributional difference in unseen tasks, on which the meta-knowledge may have less usefulness depending on the task relatedness. To overcome these limitations, we propose a novel meta-learning model that adaptively balances the effect of the meta-learning and task-specific learning within each task. Through the learning of the balancing variables, we can decide whether to obtain a solution by relying on the meta-knowledge or task-specific learning. We formulate this objective into a Bayesian inference framework and tackle it using variational inference. We validate our Bayesian Task-Adaptive Meta-Learning (Bayesian TAML) on multiple realistic task- and class-imbalanced datasets, on which it significantly outperforms existing meta-learning approaches. Further ablation study confirms the effectiveness of each balancing component and the Bayesian learning framework.-
dc.languageEnglish-
dc.publisherICLR-
dc.titleLearning to Balance: Bayesian Meta-Learning for Imbalanced and Out-of-distribution Tasks-
dc.typeConference-
dc.type.rimsCONF-
dc.citation.publicationnameEighth International Conference on Learning Representations, ICLR 2020-
dc.identifier.conferencecountryET-
dc.identifier.conferencelocationVirtual-
dc.contributor.localauthorYang, Eunho-
dc.contributor.localauthorHwang, Sung Ju-
dc.contributor.nonIdAuthorNa, Donghyun-
dc.contributor.nonIdAuthorKim, Saehoon-
dc.contributor.nonIdAuthorPark, Minseop-
Appears in Collection
RIMS Conference Papers
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0