Matches in SemOpenAlex for { <https://semopenalex.org/work/W2963203412> ?p ?o ?g. }
- W2963203412 endingPage "2149" @default.
- W2963203412 startingPage "2119" @default.
- W2963203412 abstract "We obtain a tight distribution-specific characterization of the sample complexity of large-margin classification with L2 regularization: We introduce the margin-adapted dimension, which is a simple function of the second order statistics of the data distribution, and show distribution-specific upper and lower bounds on the sample complexity, both governed by the margin-adapted dimension of the data distribution. The upper bounds are universal, and the lower bounds hold for the rich family of sub-Gaussian distributions with independent features. We conclude that this new quantity tightly characterizes the true sample complexity of large-margin classification. To prove the lower bound, we develop several new tools of independent interest. These include new connections between shattering and hardness of learning, new properties of shattering with linear classifiers, and a new lower bound on the smallest eigenvalue of a random Gram matrix generated by sub-Gaussian variables. Our results can be used to quantitatively compare large margin learning to other learning rules, and to improve the effectiveness of methods that use sample complexity bounds, such as active learning." @default.
- W2963203412 created "2019-07-30" @default.
- W2963203412 creator A5004273274 @default.
- W2963203412 creator A5070613374 @default.
- W2963203412 creator A5085310459 @default.
- W2963203412 date "2013-01-01" @default.
- W2963203412 modified "2023-09-23" @default.
- W2963203412 title "Distribution-dependent sample complexity of large margin learning" @default.
- W2963203412 cites W1482235099 @default.
- W2963203412 cites W1505356468 @default.
- W2963203412 cites W1516508599 @default.
- W2963203412 cites W1519789693 @default.
- W2963203412 cites W1542886316 @default.
- W2963203412 cites W1572210044 @default.
- W2963203412 cites W1579249934 @default.
- W2963203412 cites W1586554030 @default.
- W2963203412 cites W1589919686 @default.
- W2963203412 cites W188867022 @default.
- W2963203412 cites W1994592916 @default.
- W2963203412 cites W2004915807 @default.
- W2963203412 cites W2010029425 @default.
- W2963203412 cites W2024052500 @default.
- W2963203412 cites W2029538739 @default.
- W2963203412 cites W2042132284 @default.
- W2963203412 cites W2043856707 @default.
- W2963203412 cites W2056707879 @default.
- W2963203412 cites W2081131431 @default.
- W2963203412 cites W2081177492 @default.
- W2963203412 cites W2111377143 @default.
- W2963203412 cites W2112528242 @default.
- W2963203412 cites W2129192653 @default.
- W2963203412 cites W2144919274 @default.
- W2963203412 cites W2156909104 @default.
- W2963203412 cites W2168494960 @default.
- W2963203412 cites W2579923771 @default.
- W2963203412 cites W2951528191 @default.
- W2963203412 cites W2964121570 @default.
- W2963203412 cites W3121907121 @default.
- W2963203412 cites W3139996414 @default.
- W2963203412 cites W48137866 @default.
- W2963203412 hasPublicationYear "2013" @default.
- W2963203412 type Work @default.
- W2963203412 sameAs 2963203412 @default.
- W2963203412 citedByCount "7" @default.
- W2963203412 countsByYear W29632034122015 @default.
- W2963203412 countsByYear W29632034122018 @default.
- W2963203412 countsByYear W29632034122019 @default.
- W2963203412 countsByYear W29632034122020 @default.
- W2963203412 crossrefType "journal-article" @default.
- W2963203412 hasAuthorship W2963203412A5004273274 @default.
- W2963203412 hasAuthorship W2963203412A5070613374 @default.
- W2963203412 hasAuthorship W2963203412A5085310459 @default.
- W2963203412 hasConcept C105795698 @default.
- W2963203412 hasConcept C110121322 @default.
- W2963203412 hasConcept C11413529 @default.
- W2963203412 hasConcept C114614502 @default.
- W2963203412 hasConcept C119857082 @default.
- W2963203412 hasConcept C121332964 @default.
- W2963203412 hasConcept C12267149 @default.
- W2963203412 hasConcept C129848803 @default.
- W2963203412 hasConcept C134306372 @default.
- W2963203412 hasConcept C154945302 @default.
- W2963203412 hasConcept C158693339 @default.
- W2963203412 hasConcept C163716315 @default.
- W2963203412 hasConcept C2778445095 @default.
- W2963203412 hasConcept C2779915298 @default.
- W2963203412 hasConcept C33676613 @default.
- W2963203412 hasConcept C33923547 @default.
- W2963203412 hasConcept C41008148 @default.
- W2963203412 hasConcept C62520636 @default.
- W2963203412 hasConcept C77246614 @default.
- W2963203412 hasConcept C774472 @default.
- W2963203412 hasConcept C77553402 @default.
- W2963203412 hasConceptScore W2963203412C105795698 @default.
- W2963203412 hasConceptScore W2963203412C110121322 @default.
- W2963203412 hasConceptScore W2963203412C11413529 @default.
- W2963203412 hasConceptScore W2963203412C114614502 @default.
- W2963203412 hasConceptScore W2963203412C119857082 @default.
- W2963203412 hasConceptScore W2963203412C121332964 @default.
- W2963203412 hasConceptScore W2963203412C12267149 @default.
- W2963203412 hasConceptScore W2963203412C129848803 @default.
- W2963203412 hasConceptScore W2963203412C134306372 @default.
- W2963203412 hasConceptScore W2963203412C154945302 @default.
- W2963203412 hasConceptScore W2963203412C158693339 @default.
- W2963203412 hasConceptScore W2963203412C163716315 @default.
- W2963203412 hasConceptScore W2963203412C2778445095 @default.
- W2963203412 hasConceptScore W2963203412C2779915298 @default.
- W2963203412 hasConceptScore W2963203412C33676613 @default.
- W2963203412 hasConceptScore W2963203412C33923547 @default.
- W2963203412 hasConceptScore W2963203412C41008148 @default.
- W2963203412 hasConceptScore W2963203412C62520636 @default.
- W2963203412 hasConceptScore W2963203412C77246614 @default.
- W2963203412 hasConceptScore W2963203412C774472 @default.
- W2963203412 hasConceptScore W2963203412C77553402 @default.
- W2963203412 hasIssue "1" @default.
- W2963203412 hasOpenAccess W2963203412 @default.
- W2963203412 hasRelatedWork W1705406430 @default.
- W2963203412 hasRelatedWork W1904555434 @default.