Matches in SemOpenAlex for { <https://semopenalex.org/work/W3097700865> ?p ?o ?g. }
Showing items 1 to 77 of
77
with 100 items per page.
- W3097700865 abstract "We consider a fixed-design linear regression in the meta-learning model of Baxter (2000) and establish a problem-dependent finite-sample lower bound on the transfer risk (risk on a newly observed task) valid for all estimators. Moreover, we prove that a weighted form of a biased regularization - a popular technique in transfer and meta-learning - is optimal, i.e. it enjoys a problem-dependent upper bound on the risk matching our lower bound up to a constant. Thus, our bounds characterize meta-learning linear regression problems and reveal a fine-grained dependency on the task structure. Our characterization suggests that in the non-asymptotic regime, for a sufficiently large number of tasks, meta-learning can be considerably superior to a single-task learning. Finally, we propose a practical adaptation of the optimal estimator through Expectation-Maximization procedure and show its effectiveness in series of experiments." @default.
- W3097700865 created "2020-11-09" @default.
- W3097700865 creator A5034451303 @default.
- W3097700865 creator A5041854594 @default.
- W3097700865 creator A5069856068 @default.
- W3097700865 date "2020-10-31" @default.
- W3097700865 modified "2023-10-05" @default.
- W3097700865 title "On Optimality of Meta-Learning in Fixed-Design Regression with Weighted Biased Regularization." @default.
- W3097700865 cites W3006977545 @default.
- W3097700865 cites W3007684729 @default.
- W3097700865 hasPublicationYear "2020" @default.
- W3097700865 type Work @default.
- W3097700865 sameAs 3097700865 @default.
- W3097700865 citedByCount "3" @default.
- W3097700865 countsByYear W30977008652020 @default.
- W3097700865 countsByYear W30977008652021 @default.
- W3097700865 crossrefType "posted-content" @default.
- W3097700865 hasAuthorship W3097700865A5034451303 @default.
- W3097700865 hasAuthorship W3097700865A5041854594 @default.
- W3097700865 hasAuthorship W3097700865A5069856068 @default.
- W3097700865 hasConcept C105795698 @default.
- W3097700865 hasConcept C119857082 @default.
- W3097700865 hasConcept C126255220 @default.
- W3097700865 hasConcept C134306372 @default.
- W3097700865 hasConcept C150899416 @default.
- W3097700865 hasConcept C154945302 @default.
- W3097700865 hasConcept C165064840 @default.
- W3097700865 hasConcept C185429906 @default.
- W3097700865 hasConcept C2776135515 @default.
- W3097700865 hasConcept C2776330181 @default.
- W3097700865 hasConcept C28826006 @default.
- W3097700865 hasConcept C33923547 @default.
- W3097700865 hasConcept C41008148 @default.
- W3097700865 hasConcept C48921125 @default.
- W3097700865 hasConcept C77553402 @default.
- W3097700865 hasConceptScore W3097700865C105795698 @default.
- W3097700865 hasConceptScore W3097700865C119857082 @default.
- W3097700865 hasConceptScore W3097700865C126255220 @default.
- W3097700865 hasConceptScore W3097700865C134306372 @default.
- W3097700865 hasConceptScore W3097700865C150899416 @default.
- W3097700865 hasConceptScore W3097700865C154945302 @default.
- W3097700865 hasConceptScore W3097700865C165064840 @default.
- W3097700865 hasConceptScore W3097700865C185429906 @default.
- W3097700865 hasConceptScore W3097700865C2776135515 @default.
- W3097700865 hasConceptScore W3097700865C2776330181 @default.
- W3097700865 hasConceptScore W3097700865C28826006 @default.
- W3097700865 hasConceptScore W3097700865C33923547 @default.
- W3097700865 hasConceptScore W3097700865C41008148 @default.
- W3097700865 hasConceptScore W3097700865C48921125 @default.
- W3097700865 hasConceptScore W3097700865C77553402 @default.
- W3097700865 hasLocation W30977008651 @default.
- W3097700865 hasOpenAccess W3097700865 @default.
- W3097700865 hasPrimaryLocation W30977008651 @default.
- W3097700865 hasRelatedWork W1800306869 @default.
- W3097700865 hasRelatedWork W1914062059 @default.
- W3097700865 hasRelatedWork W2003045318 @default.
- W3097700865 hasRelatedWork W2460675832 @default.
- W3097700865 hasRelatedWork W2803642127 @default.
- W3097700865 hasRelatedWork W2890870050 @default.
- W3097700865 hasRelatedWork W2944335570 @default.
- W3097700865 hasRelatedWork W2963007973 @default.
- W3097700865 hasRelatedWork W2963252079 @default.
- W3097700865 hasRelatedWork W2963566219 @default.
- W3097700865 hasRelatedWork W3098285768 @default.
- W3097700865 hasRelatedWork W3099410262 @default.
- W3097700865 hasRelatedWork W3130990030 @default.
- W3097700865 hasRelatedWork W3132304818 @default.
- W3097700865 hasRelatedWork W3135838158 @default.
- W3097700865 hasRelatedWork W3171606515 @default.
- W3097700865 hasRelatedWork W3174145161 @default.
- W3097700865 hasRelatedWork W3199328447 @default.
- W3097700865 hasRelatedWork W3209359404 @default.
- W3097700865 hasRelatedWork W3210165725 @default.
- W3097700865 isParatext "false" @default.
- W3097700865 isRetracted "false" @default.
- W3097700865 magId "3097700865" @default.
- W3097700865 workType "article" @default.