FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2963668836> ?p ?o ?g. }

• W2963668836 abstract "Many interesting machine learning problems are best posed by considering instances that are distributions, or sample sets drawn from distributions. Most previous work devoted to machine learning tasks with distributional inputs has done so through pairwise kernel evaluations between pdfs (or sample sets). While such an approach is fine for smaller datasets, the computation of an N × N Gram matrix is prohibitive in large datasets. Recent scalable estimators that work over pdfs have done so only with kernels that use Euclidean metrics, like the L2 distance. However, there are a myriad of other useful metrics available, such as total variation, Hellinger distance, and the Jensen-Shannon divergence. This work develops the first random features for pdfs whose dot product approximates kernels using these non-Euclidean metrics. These random features allow estimators to scale to large datasets by working in a primal space, without computing large Gram matrices. We provide an analysis of the approximation error in using our proposed random features, and show empirically the quality of our approximation both in estimating a Gram matrix and in solving learning tasks in real-world and synthetic data." @default.
• W2963668836 created "2019-07-30" @default.
• W2963668836 creator A5005191773 @default.
• W2963668836 creator A5013695358 @default.
• W2963668836 creator A5050408389 @default.
• W2963668836 creator A5055199976 @default.
• W2963668836 date "2016-03-02" @default.
• W2963668836 modified "2023-10-09" @default.
• W2963668836 title "Linear-Time Learning on Distributions with Approximate Kernel Embeddings" @default.
• W2963668836 cites W1484228140 @default.
• W2963668836 cites W1587670653 @default.
• W2963668836 cites W1673728053 @default.
• W2963668836 cites W1750591676 @default.
• W2963668836 cites W1751437809 @default.
• W2963668836 cites W1780422458 @default.
• W2963668836 cites W1999764898 @default.
• W2963668836 cites W2005560667 @default.
• W2963668836 cites W2020733616 @default.
• W2963668836 cites W2058815839 @default.
• W2963668836 cites W2066941820 @default.
• W2963668836 cites W2073020428 @default.
• W2963668836 cites W2086106673 @default.
• W2963668836 cites W2098770944 @default.
• W2963668836 cites W2107325234 @default.
• W2963668836 cites W2110176078 @default.
• W2963668836 cites W2115124532 @default.
• W2963668836 cites W2115933183 @default.
• W2963668836 cites W2118585731 @default.
• W2963668836 cites W2134670479 @default.
• W2963668836 cites W2139140682 @default.
• W2963668836 cites W2144902422 @default.
• W2963668836 cites W2150879893 @default.
• W2963668836 cites W2151531457 @default.
• W2963668836 cites W2153635508 @default.
• W2963668836 cites W2157041524 @default.
• W2963668836 cites W2161690964 @default.
• W2963668836 cites W2162915993 @default.
• W2963668836 cites W2166473218 @default.
• W2963668836 cites W2168175751 @default.
• W2963668836 cites W2177692090 @default.
• W2963668836 cites W2541822344 @default.
• W2963668836 cites W2586628821 @default.
• W2963668836 cites W2737193326 @default.
• W2963668836 cites W2952400611 @default.
• W2963668836 cites W2962835968 @default.
• W2963668836 cites W2963977107 @default.
• W2963668836 cites W2964057488 @default.
• W2963668836 cites W3099652024 @default.
• W2963668836 cites W3118608800 @default.
• W2963668836 cites W569770330 @default.
• W2963668836 doi "https://doi.org/10.1609/aaai.v30i1.10308" @default.
• W2963668836 hasPublicationYear "2016" @default.
• W2963668836 type Work @default.
• W2963668836 sameAs 2963668836 @default.
• W2963668836 citedByCount "5" @default.
• W2963668836 countsByYear W29636688362016 @default.
• W2963668836 countsByYear W29636688362017 @default.
• W2963668836 countsByYear W29636688362018 @default.
• W2963668836 countsByYear W29636688362022 @default.
• W2963668836 crossrefType "journal-article" @default.
• W2963668836 hasAuthorship W2963668836A5005191773 @default.
• W2963668836 hasAuthorship W2963668836A5013695358 @default.
• W2963668836 hasAuthorship W2963668836A5050408389 @default.
• W2963668836 hasAuthorship W2963668836A5055199976 @default.
• W2963668836 hasBestOaLocation W29636688362 @default.
• W2963668836 hasConcept C105795698 @default.
• W2963668836 hasConcept C11413529 @default.
• W2963668836 hasConcept C114614502 @default.
• W2963668836 hasConcept C120174047 @default.
• W2963668836 hasConcept C121332964 @default.
• W2963668836 hasConcept C138885662 @default.
• W2963668836 hasConcept C153024298 @default.
• W2963668836 hasConcept C154945302 @default.
• W2963668836 hasConcept C158693339 @default.
• W2963668836 hasConcept C162324750 @default.
• W2963668836 hasConcept C176217482 @default.
• W2963668836 hasConcept C185429906 @default.
• W2963668836 hasConcept C186450821 @default.
• W2963668836 hasConcept C202444582 @default.
• W2963668836 hasConcept C207390915 @default.
• W2963668836 hasConcept C21547014 @default.
• W2963668836 hasConcept C28826006 @default.
• W2963668836 hasConcept C33923547 @default.
• W2963668836 hasConcept C41008148 @default.
• W2963668836 hasConcept C41895202 @default.
• W2963668836 hasConcept C62520636 @default.
• W2963668836 hasConcept C74193536 @default.
• W2963668836 hasConcept C77246614 @default.
• W2963668836 hasConceptScore W2963668836C105795698 @default.
• W2963668836 hasConceptScore W2963668836C11413529 @default.
• W2963668836 hasConceptScore W2963668836C114614502 @default.
• W2963668836 hasConceptScore W2963668836C120174047 @default.
• W2963668836 hasConceptScore W2963668836C121332964 @default.
• W2963668836 hasConceptScore W2963668836C138885662 @default.
• W2963668836 hasConceptScore W2963668836C153024298 @default.
• W2963668836 hasConceptScore W2963668836C154945302 @default.
• W2963668836 hasConceptScore W2963668836C158693339 @default.
• W2963668836 hasConceptScore W2963668836C162324750 @default.
• W2963668836 hasConceptScore W2963668836C176217482 @default.
• W2963668836 hasConceptScore W2963668836C185429906 @default.

• next