Matches in SemOpenAlex for { <https://semopenalex.org/work/W2103164654> ?p ?o ?g. }
- W2103164654 endingPage "94" @default.
- W2103164654 startingPage "79" @default.
- W2103164654 abstract "Kernel functions are typically viewed as providing an implicit mapping of points into a high-dimensional space, with the ability to gain much of the power of that space without incurring a high cost if the result is linearly-separable by a large margin γ. However, the Johnson-Lindenstrauss lemma suggests that in the presence of a large margin, a kernel function can also be viewed as a mapping to a low-dimensional space, one of dimension only $$tilde{O}(1/gamma^2)$$ . In this paper, we explore the question of whether one can efficiently produce such low-dimensional mappings, using only black-box access to a kernel function. That is, given just a program that computes K(x,y) on inputs x,y of our choosing, can we efficiently construct an explicit (small) set of features that effectively capture the power of the implicit high-dimensional space? We answer this question in the affirmative if our method is also allowed black-box access to the underlying data distribution (i.e., unlabeled examples). We also give a lower bound, showing that if we do not have access to the distribution, then this is not possible for an arbitrary black-box kernel function; we leave as an open problem, however, whether this can be done for standard kernel functions such as the polynomial kernel. Our positive result can be viewed as saying that designing a good kernel function is much like designing a good feature space. Given a kernel, by running it in a black-box manner on random unlabeled examples, we can efficiently generate an explicit set of $$tilde{O}(1/gamma^2)$$ features, such that if the data was linearly separable with margin γ under the kernel, then it is approximately separable in this new feature space." @default.
- W2103164654 created "2016-06-24" @default.
- W2103164654 creator A5024803432 @default.
- W2103164654 creator A5068544954 @default.
- W2103164654 creator A5085424042 @default.
- W2103164654 date "2006-05-03" @default.
- W2103164654 modified "2023-10-18" @default.
- W2103164654 title "Kernels as features: On kernels, margins, and low-dimensional mappings" @default.
- W2103164654 cites W1510073064 @default.
- W2103164654 cites W1542652324 @default.
- W2103164654 cites W2037757210 @default.
- W2103164654 cites W2087347434 @default.
- W2103164654 cites W2106491486 @default.
- W2103164654 cites W2108995755 @default.
- W2103164654 cites W2117362057 @default.
- W2103164654 cites W2147717514 @default.
- W2103164654 cites W2979473749 @default.
- W2103164654 cites W4249560439 @default.
- W2103164654 cites W4249572517 @default.
- W2103164654 doi "https://doi.org/10.1007/s10994-006-7550-1" @default.
- W2103164654 hasPublicationYear "2006" @default.
- W2103164654 type Work @default.
- W2103164654 sameAs 2103164654 @default.
- W2103164654 citedByCount "122" @default.
- W2103164654 countsByYear W21031646542012 @default.
- W2103164654 countsByYear W21031646542013 @default.
- W2103164654 countsByYear W21031646542014 @default.
- W2103164654 countsByYear W21031646542015 @default.
- W2103164654 countsByYear W21031646542016 @default.
- W2103164654 countsByYear W21031646542017 @default.
- W2103164654 countsByYear W21031646542018 @default.
- W2103164654 countsByYear W21031646542019 @default.
- W2103164654 countsByYear W21031646542020 @default.
- W2103164654 countsByYear W21031646542021 @default.
- W2103164654 countsByYear W21031646542022 @default.
- W2103164654 crossrefType "journal-article" @default.
- W2103164654 hasAuthorship W2103164654A5024803432 @default.
- W2103164654 hasAuthorship W2103164654A5068544954 @default.
- W2103164654 hasAuthorship W2103164654A5085424042 @default.
- W2103164654 hasBestOaLocation W21031646541 @default.
- W2103164654 hasConcept C111919701 @default.
- W2103164654 hasConcept C114614502 @default.
- W2103164654 hasConcept C118615104 @default.
- W2103164654 hasConcept C119857082 @default.
- W2103164654 hasConcept C122280245 @default.
- W2103164654 hasConcept C12267149 @default.
- W2103164654 hasConcept C134306372 @default.
- W2103164654 hasConcept C134517425 @default.
- W2103164654 hasConcept C14036430 @default.
- W2103164654 hasConcept C154945302 @default.
- W2103164654 hasConcept C160446489 @default.
- W2103164654 hasConcept C18903297 @default.
- W2103164654 hasConcept C195699287 @default.
- W2103164654 hasConcept C2777759810 @default.
- W2103164654 hasConcept C2778572836 @default.
- W2103164654 hasConcept C33676613 @default.
- W2103164654 hasConcept C33923547 @default.
- W2103164654 hasConcept C41008148 @default.
- W2103164654 hasConcept C46757340 @default.
- W2103164654 hasConcept C55851704 @default.
- W2103164654 hasConcept C70710897 @default.
- W2103164654 hasConcept C74193536 @default.
- W2103164654 hasConcept C75866337 @default.
- W2103164654 hasConcept C774472 @default.
- W2103164654 hasConcept C78458016 @default.
- W2103164654 hasConcept C86803240 @default.
- W2103164654 hasConcept C94966114 @default.
- W2103164654 hasConceptScore W2103164654C111919701 @default.
- W2103164654 hasConceptScore W2103164654C114614502 @default.
- W2103164654 hasConceptScore W2103164654C118615104 @default.
- W2103164654 hasConceptScore W2103164654C119857082 @default.
- W2103164654 hasConceptScore W2103164654C122280245 @default.
- W2103164654 hasConceptScore W2103164654C12267149 @default.
- W2103164654 hasConceptScore W2103164654C134306372 @default.
- W2103164654 hasConceptScore W2103164654C134517425 @default.
- W2103164654 hasConceptScore W2103164654C14036430 @default.
- W2103164654 hasConceptScore W2103164654C154945302 @default.
- W2103164654 hasConceptScore W2103164654C160446489 @default.
- W2103164654 hasConceptScore W2103164654C18903297 @default.
- W2103164654 hasConceptScore W2103164654C195699287 @default.
- W2103164654 hasConceptScore W2103164654C2777759810 @default.
- W2103164654 hasConceptScore W2103164654C2778572836 @default.
- W2103164654 hasConceptScore W2103164654C33676613 @default.
- W2103164654 hasConceptScore W2103164654C33923547 @default.
- W2103164654 hasConceptScore W2103164654C41008148 @default.
- W2103164654 hasConceptScore W2103164654C46757340 @default.
- W2103164654 hasConceptScore W2103164654C55851704 @default.
- W2103164654 hasConceptScore W2103164654C70710897 @default.
- W2103164654 hasConceptScore W2103164654C74193536 @default.
- W2103164654 hasConceptScore W2103164654C75866337 @default.
- W2103164654 hasConceptScore W2103164654C774472 @default.
- W2103164654 hasConceptScore W2103164654C78458016 @default.
- W2103164654 hasConceptScore W2103164654C86803240 @default.
- W2103164654 hasConceptScore W2103164654C94966114 @default.
- W2103164654 hasIssue "1" @default.
- W2103164654 hasLocation W21031646541 @default.
- W2103164654 hasOpenAccess W2103164654 @default.
- W2103164654 hasPrimaryLocation W21031646541 @default.