Practical and optimal LSH for angular distance

Alexandr Andoni; Piotr Indyk; T.M.M. Laarhoven; Ilya Razenshteyn; Ludwig Schmidt

Practical and optimal LSH for angular distance

Alexandr Andoni, Piotr Indyk, T.M.M. Laarhoven, Ilya Razenshteyn, Ludwig Schmidt

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

290 Citaties (SciVal)

Samenvatting

We show the existence of a Locality-Sensitive Hashing (LSH) family for the angular distance that yields an approximate Near Neighbor Search algorithm with the asymptotically optimal running time exponent. Unlike earlier algorithms with this property (e.g., Spherical LSH (Andoni-Indyk-Nguyen-Razenshteyn 2014) (Andoni-Razenshteyn 2015)), our algorithm is also practical, improving upon the well-studied hyperplane LSH (Charikar 2002) in practice. We also introduce a multiprobe version of this algorithm and conduct an experimental evaluation on real and synthetic data sets.We complement the above positive results with a fine-grained lower bound for the quality of any LSH family for angular distance. Our lower bound implies that the above LSH family exhibits a trade-off between evaluation time and quality that is close to optimal for a natural class of LSH functions.

Originele taal-2	Engels
Titel	Advances in Neural Information Processing Systems (NIPS, Montreal, Canada, December 7-10, 2015)
Redacteuren	C. Cortes, N.D. Lawrence, D.D. Lee, M. Sugiyama, R. Garnett
Plaats van productie	s.l.
Uitgeverij	Curran Associates
Pagina's	1225-1233
Aantal pagina's	9
Status	Gepubliceerd - 7 dec. 2015
Evenement	29th Annual Conference on Neural Information Processing Systems, NIPS 2015 - Montreal, Canada Duur: 7 dec. 2015 → 12 dec. 2015

Publicatie series

Naam	Advances in Neutral Information Processing Systems
Volume	28

Congres

Congres	29th Annual Conference on Neural Information Processing Systems, NIPS 2015
Verkorte titel	NIPS
Land/Regio	Canada
Stad	Montreal
Periode	7/12/15 → 12/12/15

Toegang tot document

https://papers.nips.cc/paper/5893-practical-and-optimal-lsh-for-angular-distance

290 Citaties
1 Rapport

Practical and optimal LSH for angular distance
Andoni, A., Indyk, P., Laarhoven, T. M. M., Razenshteyn, I. & Schmidt, L., 2015, 21 blz. (arXiv; vol. 1509.02897 [cs.DS])
Onderzoeksoutput: Boek/rapport › Rapport › Academic

Open Access
Bestand

Citeer dit

Andoni, A., Indyk, P., Laarhoven, T. M. M., Razenshteyn, I., & Schmidt, L. (2015). Practical and optimal LSH for angular distance. In C. Cortes, N. D. Lawrence, D. D. Lee, M. Sugiyama, & R. Garnett (editors), Advances in Neural Information Processing Systems (NIPS, Montreal, Canada, December 7-10, 2015) (blz. 1225-1233). (Advances in Neutral Information Processing Systems; Vol. 28). Curran Associates. https://papers.nips.cc/paper/5893-practical-and-optimal-lsh-for-angular-distance

Andoni, Alexandr ; Indyk, Piotr ; Laarhoven, T.M.M. et al. / Practical and optimal LSH for angular distance. Advances in Neural Information Processing Systems (NIPS, Montreal, Canada, December 7-10, 2015). redacteur / C. Cortes ; N.D. Lawrence ; D.D. Lee ; M. Sugiyama ; R. Garnett. s.l. : Curran Associates, 2015. blz. 1225-1233 (Advances in Neutral Information Processing Systems).

@inproceedings{ca0892cd40b645e6a08757b55f8b1ee7,

title = "Practical and optimal LSH for angular distance",

abstract = "We show the existence of a Locality-Sensitive Hashing (LSH) family for the angular distance that yields an approximate Near Neighbor Search algorithm with the asymptotically optimal running time exponent. Unlike earlier algorithms with this property (e.g., Spherical LSH (Andoni-Indyk-Nguyen-Razenshteyn 2014) (Andoni-Razenshteyn 2015)), our algorithm is also practical, improving upon the well-studied hyperplane LSH (Charikar 2002) in practice. We also introduce a multiprobe version of this algorithm and conduct an experimental evaluation on real and synthetic data sets.We complement the above positive results with a fine-grained lower bound for the quality of any LSH family for angular distance. Our lower bound implies that the above LSH family exhibits a trade-off between evaluation time and quality that is close to optimal for a natural class of LSH functions.",

author = "Alexandr Andoni and Piotr Indyk and T.M.M. Laarhoven and Ilya Razenshteyn and Ludwig Schmidt",

year = "2015",

month = dec,

day = "7",

language = "English",

series = "Advances in Neutral Information Processing Systems",

publisher = "Curran Associates",

pages = "1225--1233",

editor = "C. Cortes and N.D. Lawrence and D.D. Lee and M. Sugiyama and R. Garnett",

booktitle = "Advances in Neural Information Processing Systems (NIPS, Montreal, Canada, December 7-10, 2015)",

note = "29th Annual Conference on Neural Information Processing Systems, NIPS 2015, NIPS ; Conference date: 07-12-2015 Through 12-12-2015",

}

Andoni, A, Indyk, P, Laarhoven, TMM, Razenshteyn, I & Schmidt, L 2015, Practical and optimal LSH for angular distance. in C Cortes, ND Lawrence, DD Lee, M Sugiyama & R Garnett (uitgave), Advances in Neural Information Processing Systems (NIPS, Montreal, Canada, December 7-10, 2015). Advances in Neutral Information Processing Systems, vol. 28, Curran Associates, s.l., blz. 1225-1233, 29th Annual Conference on Neural Information Processing Systems, NIPS 2015, Montreal, Canada, 7/12/15. <https://papers.nips.cc/paper/5893-practical-and-optimal-lsh-for-angular-distance>

Practical and optimal LSH for angular distance. / Andoni, Alexandr; Indyk, Piotr; Laarhoven, T.M.M. et al.
Advances in Neural Information Processing Systems (NIPS, Montreal, Canada, December 7-10, 2015). uitgave / C. Cortes; N.D. Lawrence; D.D. Lee; M. Sugiyama; R. Garnett. s.l.: Curran Associates, 2015. blz. 1225-1233 (Advances in Neutral Information Processing Systems; Vol. 28).

Onderzoeksoutput: Hoofdstuk in Boek/Rapport/Congresprocedure › Conferentiebijdrage › Academic › peer review

TY - GEN

T1 - Practical and optimal LSH for angular distance

AU - Andoni, Alexandr

AU - Indyk, Piotr

AU - Laarhoven, T.M.M.

AU - Razenshteyn, Ilya

AU - Schmidt, Ludwig

PY - 2015/12/7

Y1 - 2015/12/7

N2 - We show the existence of a Locality-Sensitive Hashing (LSH) family for the angular distance that yields an approximate Near Neighbor Search algorithm with the asymptotically optimal running time exponent. Unlike earlier algorithms with this property (e.g., Spherical LSH (Andoni-Indyk-Nguyen-Razenshteyn 2014) (Andoni-Razenshteyn 2015)), our algorithm is also practical, improving upon the well-studied hyperplane LSH (Charikar 2002) in practice. We also introduce a multiprobe version of this algorithm and conduct an experimental evaluation on real and synthetic data sets.We complement the above positive results with a fine-grained lower bound for the quality of any LSH family for angular distance. Our lower bound implies that the above LSH family exhibits a trade-off between evaluation time and quality that is close to optimal for a natural class of LSH functions.

AB - We show the existence of a Locality-Sensitive Hashing (LSH) family for the angular distance that yields an approximate Near Neighbor Search algorithm with the asymptotically optimal running time exponent. Unlike earlier algorithms with this property (e.g., Spherical LSH (Andoni-Indyk-Nguyen-Razenshteyn 2014) (Andoni-Razenshteyn 2015)), our algorithm is also practical, improving upon the well-studied hyperplane LSH (Charikar 2002) in practice. We also introduce a multiprobe version of this algorithm and conduct an experimental evaluation on real and synthetic data sets.We complement the above positive results with a fine-grained lower bound for the quality of any LSH family for angular distance. Our lower bound implies that the above LSH family exhibits a trade-off between evaluation time and quality that is close to optimal for a natural class of LSH functions.

M3 - Conference contribution

T3 - Advances in Neutral Information Processing Systems

SP - 1225

EP - 1233

BT - Advances in Neural Information Processing Systems (NIPS, Montreal, Canada, December 7-10, 2015)

A2 - Cortes, C.

A2 - Lawrence, N.D.

A2 - Lee, D.D.

A2 - Sugiyama, M.

A2 - Garnett, R.

PB - Curran Associates

CY - s.l.

T2 - 29th Annual Conference on Neural Information Processing Systems, NIPS 2015

Y2 - 7 December 2015 through 12 December 2015

ER -

Andoni A, Indyk P, Laarhoven TMM, Razenshteyn I, Schmidt L. Practical and optimal LSH for angular distance. In Cortes C, Lawrence ND, Lee DD, Sugiyama M, Garnett R, redacteurs, Advances in Neural Information Processing Systems (NIPS, Montreal, Canada, December 7-10, 2015). s.l.: Curran Associates. 2015. blz. 1225-1233. (Advances in Neutral Information Processing Systems).

Practical and optimal LSH for angular distance

Samenvatting

Publicatie series

Congres

Toegang tot document

Vingerafdruk

Onderzoekersoutput

Practical and optimal LSH for angular distance

Citeer dit