Abstract
Locality-sensitive hashing (LSH), introduced by Indyk and Motwani in STOC ’98, has been an extremely influential framework for nearest neighbor search in high-dimensional data sets. While theoretical work has focused on the approximate nearest neighbor problem, in practice LSH data structures with suitably chosen parameters are used to solve the exact nearest neighbor problem (with some error probability). Sublinear query time is often possible in practice even for exact nearest neighbor search, intuitively because the nearest neighbor tends to be significantly closer than other data points. However, theory offers little advice on how to choose LSH parameters outside of pre-specified worst-case settings.
We introduce the technique of confirmation sampling for solving the exact nearest neighbor problem using LSH. First, we give a general reduction that transforms a sequence of data structures that each find the nearest neighbor with a small, unknown probability, into a data structure that returns the nearest neighbor with probability \(1-\delta \), using as few queries as possible. Second, we present a new query algorithm for the LSH Forest data structure with L trees that is able to return the exact nearest neighbor of a query point within the same time bound as an LSH Forest of \(\varOmega (L)\) trees with internal parameters specifically tuned to the query and data.
T. Christiani—The research leading to these results has received funding from the European Research Council under the European Union’s 7th Framework Programme (FP7/2007-2013)/ERC grant agreement no. 614331.
R. Pagh—Supported by Villum Foundation grant 16582 to Basic Algorithms Research Copenhagen (BARC). Part of this work was done while visiting Simons Institute for the Theory of Computing.
M. Thorup—Supported by an Investigator Grant from the Villum Foundation, Grant No. 16582.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
The sampling of a random element ensures compatibility with ConfirmationSampling, which requires a sample to be returned even if there is no hash collision. It is not really necessary from an algorithmic viewpoint, but also does not hurt the asymptotic performance.
- 2.
For every choice of constant \(c \ge 1\) there exists a constant \(n_0\) such that for \(n \ge n_0\) we can obtain success probability \(1 - 1/n^c\) where \(n = |P|\) denotes the size of the set of data points.
References
Andoni, A., Indyk, P., Laarhoven, T., Razenshteyn, I., Schmidt, L.: Practical and optimal LSH for angular distance. In: Proceedings of the NIPS 2015, pp. 1225–1233 (2015)
Andoni, A., Laarhoven, T., Razenshteyn, I.P., Waingarten, E.: Optimal hashing-based time-space trade-offs for approximate near neighbors. In: Proceedings of the SODA 2017, pp. 47–66 (2017)
Andoni, A., Razenshteyn, I.: Optimal data-dependent hashing for approximate near neighbors. In: Proceedings of the STOC 2015, pp. 793–801 (2015)
Aumüller, M., Bernhardsson, E., Faithfull, A.: ANN-benchmarks: a benchmarking tool for approximate nearest neighbor algorithms. In: Beecks, C., Borutta, F., Kröger, P., Seidl, T. (eds.) SISAP 2017. LNCS, vol. 10609, pp. 34–49. Springer, Cham (2017). https://doi.org/10.1007/978-3-319-68474-1_3
Aumüller, M., Christiani, T., Pagh, R., Vesterli, M.: PUFFINN: parameterless and universally fast finding of nearest neighbors. In: Proceedings of the ESA 2019. LIPIcs, vol. 144, pp. 10:1–10:16 (2019)
Bawa, M., Condie, T., Ganesan, P.: LSH forest: self-tuning indexes for similarity search. In: Proceedings of the WWW 2005, pp. 651–660 (2005)
Charikar, M.: Similarity estimation techniques from rounding algorithms. In: Proceedings of the STOC 2002, pp. 380–388 (2002)
Christiani, T.: Fast locality-sensitive hashing frameworks for approximate near neighbor search. In: Amato, G., Gennaro, C., Oria, V., Radovanović, M. (eds.) SISAP 2019. LNCS, vol. 11807, pp. 3–17. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-32047-8_1
Christiani, T., Pagh, R., Thorup, M.: Confirmation sampling for exact nearest neighbor search. CoRR abs/1812.02603 (2018). http://arxiv.org/abs/1812.02603
Datar, M., Immorlica, N., Indyk, P., Mirrokni, V.S.: Locality-sensitive hashing scheme based on p-stable distributions. In: Proceedings of the SOCG 2004, pp. 253–262 (2004)
Dong, W., Wang, Z., Josephson, W., Charikar, M., Li, K.: Modeling LSH for performance tuning. In: Proceedings of the CIKM 2008, pp. 669–678 (2008)
Har-Peled, S., Indyk, P., Motwani, R.: Approximate nearest neighbor: towards removing the curse of dimensionality. Theor. Comput. 8(1), 321–350 (2012)
Indyk, P., Motwani, R.: Approximate nearest neighbors: towards removing the curse of dimensionality. In: Proceedings of the STOC 1998, pp. 604–613 (1998)
Li, P., König, A.C.: Theory and applications of b-bit minwise hashing. Commun. ACM 54(8), 101–109 (2011)
Lv, Q., Josephson, W., Wang, Z., Charikar, M., Li, K.: Intelligent probing for locality sensitive hashing: multi-probe LSH and beyond. PVLDB 10(12), 2021–2024 (2017)
Panigrahy, R.: Entropy based nearest neighbor search in high dimensions. In: Proceedings of the SODA 2006, pp. 1186–1195 (2006)
Slaney, M., Lifshits, Y., He, J.: Optimal parameters for locality-sensitive hashing. Proc. IEEE 100(9), 2604–2623 (2012)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Christiani, T., Pagh, R., Thorup, M. (2020). Confirmation Sampling for Exact Nearest Neighbor Search. In: Satoh, S., et al. Similarity Search and Applications. SISAP 2020. Lecture Notes in Computer Science(), vol 12440. Springer, Cham. https://doi.org/10.1007/978-3-030-60936-8_8
Download citation
DOI: https://doi.org/10.1007/978-3-030-60936-8_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-60935-1
Online ISBN: 978-3-030-60936-8
eBook Packages: Computer ScienceComputer Science (R0)