Filters








18,089 Hits in 7.3 sec

Approximate Nearest Neighbor Search on High Dimensional Data --- Experiments, Analyses, and Improvement (v1.0) [article]

Wen Li, Ying Zhang, Yifang Sun, Wei Wang, Wenjie Zhang, Xuemin Lin
2016 arXiv   pre-print
Approximate Nearest neighbor search (ANNS) is fundamental and essential operation in applications from many domains, such as databases, machine learning, multimedia, and computer vision.  ...  In this paper, we conduct a comprehensive experimental evaluation of many state-of-the-art methods for approximate nearest neighbor search.  ...  Due to the curse of dimensionality [26] , much research efforts focus on the approximate solution for the problem of k nearest neighbor search on high dimensional data.  ... 
arXiv:1610.02455v1 fatcat:skn6iftztnhr7d524fqvyqix3m

Scalable Nearest Neighbor Algorithms for High Dimensional Data

Marius Muja, David G. Lowe
2014 IEEE Transactions on Pattern Analysis and Machine Intelligence  
Index Terms-Nearest neighbor search, big data, approximate search, algorithm configuration Ç M. Muja is with BitLit  ...  We show that the optimal nearest neighbor algorithm and its parameters depend on the data set characteristics and describe an automated configuration procedure for finding the best algorithm to search  ...  Fast Approximate Nearest Neighbor Search We present several experiments we have conducted in order to analyse the performance of the two algorithms described in Section 3.  ... 
doi:10.1109/tpami.2014.2321376 pmid:26353063 fatcat:n42wwark2fdephku4jsjwfv5ky

Leveraging Reinforcement Learning for evaluating Robustness of KNN Search Algorithms [article]

Pramod Vadiraja, Christoph Peter Balada
2021 arXiv   pre-print
With the issue of curse of dimensionality, it gets quite tedious to reliably bank on the results of variety approximate nearest neighbor search approaches.  ...  In very high dimensional spaces the K-nearest neighbor search (KNNS) suffers in terms of complexity in computation of high dimensional distances.  ...  Hence the high dimensional points that failed to result in true nearest neighbors by some KNNS techniques might have to be analysed deeper.  ... 
arXiv:2102.06525v1 fatcat:d5tfa2nuknbzzcld42satedhiy

Fast and Scalable Outlier Detection with Approximate Nearest Neighbor Ensembles [chapter]

Erich Schubert, Arthur Zimek, Hans-Peter Kriegel
2015 Lecture Notes in Computer Science  
hashing and projection indexed nearest neighbors.  ...  Existing approximate neighbor search methods are designed to preserve distances as well as possible.  ...  If the data dimensionality is low, bulk-loaded R*-trees are excellent. 2. If the exact distances are of importance, PINN is expected to work best. 3.  ... 
doi:10.1007/978-3-319-18123-3_2 fatcat:jbhn3t6vhrbqdcwwnjdzhaz3zu

Revisiting k-Nearest Neighbor Graph Construction on High-Dimensional Data : Experiments and Analyses [article]

Liu Yingfan, Cheng Hong, Cui Jiangtao
2021 arXiv   pre-print
The k-nearest neighbor graph (KNNG) on high-dimensional data is a data structure widely used in many applications such as similarity search, dimension reduction and clustering.  ...  To address these issues, we comprehensively compare the representative approaches on real-world high-dimensional data sets to provide practical and insightful suggestions for users.  ...  – experiments, analyses, and improvement.  ... 
arXiv:2112.02234v1 fatcat:od4skj4worgvfc7dbonwof7y2y

Connecting Compression Spaces with Transformer for Approximate Nearest Neighbor Search [article]

Haokui Zhang, Buzhou Tang, Wenze Hu, Xiaoyu Wang
2022 arXiv   pre-print
We propose a generic feature compression method for Approximate Nearest Neighbor Search (ANNS) problems, which speeds up existing ANNS methods in a plug-and-play manner.  ...  to maintain high search accuracy.  ...  INTRODUCTION Approximate nearest neighbor search (ANNS) methods focus on searching for k approximate nearest neighbors from a given database to a given query node q.  ... 
arXiv:2107.14415v5 fatcat:e6mh4cg53bh77h5fla6inp2zia

Breaking the curse of dimensionality with Isolation Kernel [article]

Kai Ming Ting, Takashi Washio, Ye Zhu, Yang Xu
2021 arXiv   pre-print
We show that only Isolation Kernel performs consistently well in indexed search, spectral & density peaks clustering, SVM classification and t-SNE visualization in both low and high dimensions, compared  ...  The curse of dimensionality has been studied in different aspects. However, breaking the curse has been elusive.  ...  than brute force search and have equally high precision on this low dimensional dataset.  ... 
arXiv:2109.14198v1 fatcat:hilkssihu5aevi6vp2mzvh7o3m

NV-Tree

Herwig Lejsek, Björn Þór Jónsson, Laurent Amsaleg
2011 Proceedings of the 1st ACM International Conference on Multimedia Retrieval - ICMR '11  
It addresses the specific, yet important, problem of efficiently and effectively finding the approximate k-nearest neighbors within a collection of a few billion high-dimensional data points.  ...  Large Scale High-Dimensional Data Sets Recently, various approaches for k-nn retrieval at large scale were proposed, some even termed as addressing "webscale problems" [1, 17, 3] .  ...  of few billion data points stored on disk, the approximate k-nearest neighbors of individual query points.  ... 
doi:10.1145/1991996.1992050 dblp:conf/mir/LejsekJA11 fatcat:hesva7eem5emfivgtxxtyfp3cy

Fast protein 3D surface search

Sungchul Kim, Lee Sael, Hwanjo Yu
2013 Proceedings of the 7th International Conference on Ubiquitous Information Management and Communication - ICUIMC '13  
Experiments show that the searching time reduced 75.41% by the fast k-nearest neighbor algorithm, 88.7% by the extended fast k-nearest neighbor algorithm, 88.84% by the fast threshold-based nearest neighbor  ...  algorithm, and 91.53% by the fast extended threshold-based nearest neighbor algorithm.  ...  In this work, we exploit the fast nearest neighbor search introduced by Hwang et al [8] which can efficiently retrieve exact nearest neighbors on high-dimensional data.  ... 
doi:10.1145/2448556.2448629 dblp:conf/icuimc/KimSY13 fatcat:ouzhrmunurbxfjzmx7ql5vdxwm

Efficient k-nearest neighbor searches for multi-source forest attribute mapping

Andrew O. Finley, Ronald E. McRoberts
2008 Remote Sensing of Environment  
Further, given our trial data, we found that enormous gain in search time efficiency, afforded by approximate nearest neighbor search algorithms, does not result in compromised kNN prediction.  ...  We conclude that by using the kd-tree, or similar data structure, and efficient exact or approximate search algorithms, the kNN method, and variants, are useful tools for mapping large geographic areas  ...  The focus of this paper is on the practical application of specialized data structures that optimally partition the d-dimensional space to facilitate efficient nearest neighbor searches.  ... 
doi:10.1016/j.rse.2007.08.024 fatcat:llgrkpcizfha5j6u42mct76tem

A Distributed and Approximated Nearest Neighbors Algorithm for an Efficient Large Scale Mean Shift Clustering [article]

Gaël Beck, Tarn Duong, Mustapha Lebbah, Hanane Azzag, Christophe Cérin
2019 arXiv   pre-print
Both propositions are based on Locality Sensitive Hashing (LSH) to approximate nearest neighbors. These two techniques may be used for moderate sized datasets.  ...  Furthermore, we show that using our proposed approximations of the density gradient ascent as a pre-processing step in other clustering methods can also improve dedicated classification metrics.  ...  Acknowledgments Experiments presented in this paper were carried out using the Grid'5000 testbed, supported by a scientific interest group hosted by Inria and including CNRS, RENATER and several Universities  ... 
arXiv:1902.03833v1 fatcat:alham5m6rzbrnpmdqbimggedgm

MLR-Index: An Index Structure for Fast and Scalable Similarity Search in High Dimensions [chapter]

Rahul Malik, Sangkyum Kim, Xin Jin, Chandrasekar Ramachandran, Jiawei Han, Indranil Gupta, Klara Nahrstedt
2009 Lecture Notes in Computer Science  
In this paper, we focus on an approximate nearest neighbor search for two different types of queries: r-Range search and k-NN search.  ...  Because of the curse of dimensionality, it is already known that well-known data structures like kd-tree, R-tree, and M-tree suffer in their performance over high-dimensional data space which is inferior  ...  -We propose approximate algorithms which are time and space efficient with high accuracy for two different nearest neighbor search problems.  ... 
doi:10.1007/978-3-642-02279-1_12 fatcat:gmaphojg7zbc3jp5xesn6fo6cm

Tree-based Search Graph for Approximate Nearest Neighbor Search [article]

Xiaobin Fan, Xiaoping Wang, Kai Lu, Lei Xue, Jinjing Zhao
2022 arXiv   pre-print
Since the computational cost for accurate search is too high, the community turned to the research of approximate nearest neighbor search (ANNS).  ...  Nearest neighbor search supports important applications in many domains, such as database, machine learning, computer vision.  ...  However, exact nearest neighbor search in high dimensional space is often computationally expensive.  ... 
arXiv:2201.03237v1 fatcat:53wv6jw4tbamljcsl2cedbuysq

Graph-Based Time-Space Trade-Offs for Approximate Near Neighbors

Thijs Laarhoven, Marc Herbstritt
2018 International Symposium on Computational Geometry  
walks on the approximate nearest neighbor graph.  ...  For random data sets of size n = 2 o(d) on the d-dimensional Euclidean unit sphere, using near neighbor graphs we can provably solve the approximate nearest neighbor problem with approximation factor c  ...  methods for nearest neighbor searching on various commonly used data sets and distance metrics [13, 9] .  ... 
doi:10.4230/lipics.socg.2018.57 dblp:conf/compgeom/Laarhoven18 fatcat:ybdrcak45vbirce44fbnwu6hyq

Angle Tree: Nearest Neighbor Search in High Dimensions with Low Intrinsic Dimensionality [article]

Ilia Zvedeniouk, Sanjay Chawla
2010 arXiv   pre-print
Experiments and analysis on real and synthetic data sets shows that the Angle Tree is the most efficient known indexing structure for nearest neighbor queries in terms of preprocessing and space usage  ...  while achieving high accuracy and fast search time.  ...  [1] introduced a novel approach to analysing low dimensional manifolds embedded in high dimensions.  ... 
arXiv:1003.5474v2 fatcat:iwmrw72ayndadk6ke3c2vpjlxi
« Previous Showing results 1 — 15 out of 18,089 results