FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W1521083607> ?p ?o ?g. }

• W1521083607 endingPage "133" @default.
• W1521083607 startingPage "119" @default.
• W1521083607 abstract "In this paper we prove various results about PAC learning in the presence of malicious and random classification noise. Our main theme is the use of randomized hypotheses for learning with small sample sizes and high malicious noise rates. We show an algorithm that PAC learns any target class of VC-dimension d using randomized hypotheses and order of d/ε training examples (up to logarithmic factors) while tolerating malicious noise rates even slightly larger than the information-theoretic bound ε/(1+ε) for deterministic hypotheses. Combined with previous results, this implies that a lower bound d/Δ+ε/Δ 2 on the sample size, where η=ε/(l+ε)−Δ is the malicious noise rate, applies only when using deterministic hypotheses. We then show that the information-theoretic upper bound on the noise rate for deterministic hypotheses can be replaced by 2ε/(l+2ε) if randomized hypotheses are used. Investigating further the use of randomized hypotheses, we show a strategy for learning the powerset of d elements using an optimal sample size of order dε/Δ 2 (up to logarithmic factors) and tolerating a noise rate η=2ε/(l+2ε)−Δ. We complement this result by proving that this sample size is also necessary for any class C of VC-dimension d.We then discuss the performance of the minimum disagreement strategy under both malicious and random classification noise models. For malicious noise we show an algorithm that, using deterministic hypotheses, learns unions of d intervals on the continuous domain [0, 1) using a sample size significantly smaller than that needed by the minimum disagreement strategy. For classification noise we show, generalizing a result by Laird, that order of d/εΔ 2) training examples suffice (up to logarithmic factors) to learn by minimizing disagreements any target class of VC-dimension d tolerating random classification noise rate η=1/2−Δ. Using a lower bound by Simon, we also prove that this sample size bound cannot be significantly improved." @default.
• W1521083607 created "2016-06-24" @default.
• W1521083607 creator A5032272542 @default.
• W1521083607 creator A5033310382 @default.
• W1521083607 creator A5056190255 @default.
• W1521083607 creator A5074621968 @default.
• W1521083607 date "1997-01-01" @default.
• W1521083607 modified "2023-09-28" @default.
• W1521083607 title "Randomized hypotheses and minimum disagreement hypotheses for learning with noise" @default.
• W1521083607 cites W1507495319 @default.
• W1521083607 cites W1968998685 @default.
• W1521083607 cites W2045313701 @default.
• W1521083607 cites W2129192653 @default.
• W1521083607 cites W2149483545 @default.
• W1521083607 cites W4232150823 @default.
• W1521083607 cites W4238893454 @default.
• W1521083607 doi "https://doi.org/10.1007/3-540-62685-9_11" @default.
• W1521083607 hasPublicationYear "1997" @default.
• W1521083607 type Work @default.
• W1521083607 sameAs 1521083607 @default.
• W1521083607 citedByCount "4" @default.
• W1521083607 countsByYear W15210836072013 @default.
• W1521083607 countsByYear W15210836072014 @default.
• W1521083607 countsByYear W15210836072016 @default.
• W1521083607 crossrefType "book-chapter" @default.
• W1521083607 hasAuthorship W1521083607A5032272542 @default.
• W1521083607 hasAuthorship W1521083607A5033310382 @default.
• W1521083607 hasAuthorship W1521083607A5056190255 @default.
• W1521083607 hasAuthorship W1521083607A5074621968 @default.
• W1521083607 hasConcept C104317684 @default.
• W1521083607 hasConcept C105795698 @default.
• W1521083607 hasConcept C112313634 @default.
• W1521083607 hasConcept C11413529 @default.
• W1521083607 hasConcept C114614502 @default.
• W1521083607 hasConcept C115961682 @default.
• W1521083607 hasConcept C127716648 @default.
• W1521083607 hasConcept C128669082 @default.
• W1521083607 hasConcept C129848803 @default.
• W1521083607 hasConcept C134306372 @default.
• W1521083607 hasConcept C154945302 @default.
• W1521083607 hasConcept C185592680 @default.
• W1521083607 hasConcept C188082640 @default.
• W1521083607 hasConcept C198531522 @default.
• W1521083607 hasConcept C203313322 @default.
• W1521083607 hasConcept C2777212361 @default.
• W1521083607 hasConcept C33676613 @default.
• W1521083607 hasConcept C33923547 @default.
• W1521083607 hasConcept C39927690 @default.
• W1521083607 hasConcept C41008148 @default.
• W1521083607 hasConcept C43617362 @default.
• W1521083607 hasConcept C55493867 @default.
• W1521083607 hasConcept C77553402 @default.
• W1521083607 hasConcept C99498987 @default.
• W1521083607 hasConceptScore W1521083607C104317684 @default.
• W1521083607 hasConceptScore W1521083607C105795698 @default.
• W1521083607 hasConceptScore W1521083607C112313634 @default.
• W1521083607 hasConceptScore W1521083607C11413529 @default.
• W1521083607 hasConceptScore W1521083607C114614502 @default.
• W1521083607 hasConceptScore W1521083607C115961682 @default.
• W1521083607 hasConceptScore W1521083607C127716648 @default.
• W1521083607 hasConceptScore W1521083607C128669082 @default.
• W1521083607 hasConceptScore W1521083607C129848803 @default.
• W1521083607 hasConceptScore W1521083607C134306372 @default.
• W1521083607 hasConceptScore W1521083607C154945302 @default.
• W1521083607 hasConceptScore W1521083607C185592680 @default.
• W1521083607 hasConceptScore W1521083607C188082640 @default.
• W1521083607 hasConceptScore W1521083607C198531522 @default.
• W1521083607 hasConceptScore W1521083607C203313322 @default.
• W1521083607 hasConceptScore W1521083607C2777212361 @default.
• W1521083607 hasConceptScore W1521083607C33676613 @default.
• W1521083607 hasConceptScore W1521083607C33923547 @default.
• W1521083607 hasConceptScore W1521083607C39927690 @default.
• W1521083607 hasConceptScore W1521083607C41008148 @default.
• W1521083607 hasConceptScore W1521083607C43617362 @default.
• W1521083607 hasConceptScore W1521083607C55493867 @default.
• W1521083607 hasConceptScore W1521083607C77553402 @default.
• W1521083607 hasConceptScore W1521083607C99498987 @default.
• W1521083607 hasLocation W15210836071 @default.
• W1521083607 hasOpenAccess W1521083607 @default.
• W1521083607 hasPrimaryLocation W15210836071 @default.
• W1521083607 hasRelatedWork W1539705875 @default.
• W1521083607 hasRelatedWork W1813877927 @default.
• W1521083607 hasRelatedWork W1849101968 @default.
• W1521083607 hasRelatedWork W1977696979 @default.
• W1521083607 hasRelatedWork W1981517317 @default.
• W1521083607 hasRelatedWork W1991417304 @default.
• W1521083607 hasRelatedWork W2005530305 @default.
• W1521083607 hasRelatedWork W2019388495 @default.
• W1521083607 hasRelatedWork W2110001975 @default.
• W1521083607 hasRelatedWork W2114949160 @default.
• W1521083607 hasRelatedWork W2247708794 @default.
• W1521083607 hasRelatedWork W2255026220 @default.
• W1521083607 hasRelatedWork W2509517678 @default.
• W1521083607 hasRelatedWork W2607437843 @default.
• W1521083607 hasRelatedWork W2941363245 @default.
• W1521083607 hasRelatedWork W2963007973 @default.
• W1521083607 hasRelatedWork W2963103083 @default.
• W1521083607 hasRelatedWork W2964242659 @default.

• next