Matches in SemOpenAlex for { <https://semopenalex.org/work/W2921052765> ?p ?o ?g. }
- W2921052765 abstract "Feature hashing and other random projection schemes are commonly used to reduce the dimensionality of feature vectors. The goal is to efficiently project a high-dimensional feature vector living in $mathbb{R}^n$ into a much lower-dimensional space $mathbb{R}^m$, while approximately preserving Euclidean norm. These schemes can be constructed using sparse random projections, for example using a sparse Johnson-Lindenstrauss (JL) transform. A line of work introduced by Weinberger et. al (ICML '09) analyzes the accuracy of sparse JL with sparsity 1 on feature vectors with small $ell_infty$-to-$ell_2$ norm ratio. Recently, Freksen, Kamma, and Larsen (NeurIPS '18) closed this line of work by proving a tight tradeoff between $ell_infty$-to-$ell_2$ norm ratio and accuracy for sparse JL with sparsity $1$. In this paper, we demonstrate the benefits of using sparsity $s$ greater than $1$ in sparse JL on feature vectors. Our main result is a tight tradeoff between $ell_infty$-to-$ell_2$ norm ratio and accuracy for a general sparsity $s$, that significantly generalizes the result of Freksen et. al. Our result theoretically demonstrates that sparse JL with $s > 1$ can have significantly better norm-preservation properties on feature vectors than sparse JL with $s = 1$; we also empirically demonstrate this finding." @default.
- W2921052765 created "2019-03-22" @default.
- W2921052765 creator A5005798722 @default.
- W2921052765 date "2019-03-08" @default.
- W2921052765 modified "2023-09-23" @default.
- W2921052765 title "Understanding Sparse JL for Feature Hashing" @default.
- W2921052765 cites W1000140482 @default.
- W2921052765 cites W1746783006 @default.
- W2921052765 cites W1794558985 @default.
- W2921052765 cites W1989591518 @default.
- W2921052765 cites W2012362409 @default.
- W2921052765 cites W2021970885 @default.
- W2921052765 cites W2025423507 @default.
- W2921052765 cites W2037757210 @default.
- W2921052765 cites W2041836310 @default.
- W2921052765 cites W2052805042 @default.
- W2921052765 cites W2070996757 @default.
- W2921052765 cites W2072864444 @default.
- W2921052765 cites W2085927849 @default.
- W2921052765 cites W2114773674 @default.
- W2921052765 cites W2121689290 @default.
- W2921052765 cites W2132032691 @default.
- W2921052765 cites W2135132414 @default.
- W2921052765 cites W2172166488 @default.
- W2921052765 cites W2295512376 @default.
- W2921052765 cites W2570762126 @default.
- W2921052765 cites W2767577269 @default.
- W2921052765 cites W2783154142 @default.
- W2921052765 cites W2792300954 @default.
- W2921052765 cites W2885241647 @default.
- W2921052765 cites W2962766560 @default.
- W2921052765 cites W2964064010 @default.
- W2921052765 cites W2977251998 @default.
- W2921052765 cites W2979473749 @default.
- W2921052765 hasPublicationYear "2019" @default.
- W2921052765 type Work @default.
- W2921052765 sameAs 2921052765 @default.
- W2921052765 citedByCount "5" @default.
- W2921052765 countsByYear W29210527652020 @default.
- W2921052765 countsByYear W29210527652021 @default.
- W2921052765 crossrefType "posted-content" @default.
- W2921052765 hasAuthorship W2921052765A5005798722 @default.
- W2921052765 hasConcept C105795698 @default.
- W2921052765 hasConcept C111030470 @default.
- W2921052765 hasConcept C11413529 @default.
- W2921052765 hasConcept C114614502 @default.
- W2921052765 hasConcept C118615104 @default.
- W2921052765 hasConcept C124851039 @default.
- W2921052765 hasConcept C129782007 @default.
- W2921052765 hasConcept C133667856 @default.
- W2921052765 hasConcept C138111711 @default.
- W2921052765 hasConcept C138885662 @default.
- W2921052765 hasConcept C153180895 @default.
- W2921052765 hasConcept C154945302 @default.
- W2921052765 hasConcept C17744445 @default.
- W2921052765 hasConcept C178489894 @default.
- W2921052765 hasConcept C191795146 @default.
- W2921052765 hasConcept C199539241 @default.
- W2921052765 hasConcept C2524010 @default.
- W2921052765 hasConcept C2776401178 @default.
- W2921052765 hasConcept C2777036070 @default.
- W2921052765 hasConcept C33923547 @default.
- W2921052765 hasConcept C41008148 @default.
- W2921052765 hasConcept C41895202 @default.
- W2921052765 hasConcept C7608002 @default.
- W2921052765 hasConcept C83665646 @default.
- W2921052765 hasConceptScore W2921052765C105795698 @default.
- W2921052765 hasConceptScore W2921052765C111030470 @default.
- W2921052765 hasConceptScore W2921052765C11413529 @default.
- W2921052765 hasConceptScore W2921052765C114614502 @default.
- W2921052765 hasConceptScore W2921052765C118615104 @default.
- W2921052765 hasConceptScore W2921052765C124851039 @default.
- W2921052765 hasConceptScore W2921052765C129782007 @default.
- W2921052765 hasConceptScore W2921052765C133667856 @default.
- W2921052765 hasConceptScore W2921052765C138111711 @default.
- W2921052765 hasConceptScore W2921052765C138885662 @default.
- W2921052765 hasConceptScore W2921052765C153180895 @default.
- W2921052765 hasConceptScore W2921052765C154945302 @default.
- W2921052765 hasConceptScore W2921052765C17744445 @default.
- W2921052765 hasConceptScore W2921052765C178489894 @default.
- W2921052765 hasConceptScore W2921052765C191795146 @default.
- W2921052765 hasConceptScore W2921052765C199539241 @default.
- W2921052765 hasConceptScore W2921052765C2524010 @default.
- W2921052765 hasConceptScore W2921052765C2776401178 @default.
- W2921052765 hasConceptScore W2921052765C2777036070 @default.
- W2921052765 hasConceptScore W2921052765C33923547 @default.
- W2921052765 hasConceptScore W2921052765C41008148 @default.
- W2921052765 hasConceptScore W2921052765C41895202 @default.
- W2921052765 hasConceptScore W2921052765C7608002 @default.
- W2921052765 hasConceptScore W2921052765C83665646 @default.
- W2921052765 hasOpenAccess W2921052765 @default.
- W2921052765 hasRelatedWork W1488444356 @default.
- W2921052765 hasRelatedWork W1674879387 @default.
- W2921052765 hasRelatedWork W2085927849 @default.
- W2921052765 hasRelatedWork W2272355041 @default.
- W2921052765 hasRelatedWork W2465082976 @default.
- W2921052765 hasRelatedWork W2532988047 @default.
- W2921052765 hasRelatedWork W2557240553 @default.
- W2921052765 hasRelatedWork W2594606127 @default.
- W2921052765 hasRelatedWork W2757458919 @default.