66,799 Hits in 7.9 sec

Accelerating Large-Scale Inference with Anisotropic Vector Quantization [article]

Ruiqi Guo, Philip Sun, Erik Lindgren, Quan Geng, David Simcha, Felix Chern, Sanjiv Kumar
2020 arXiv   pre-print
Quantization based techniques are the current state-of-the-art for scaling maximum inner product search to massive databases.  ...  Based on the observation that for a given query, the database points that have the largest inner products are more relevant, we develop a family of anisotropic quantization loss functions.  ...  Conclusion In this paper, we propose a new quantization loss function for inner product search, which replaces traditional reconstruction error.  ... 
arXiv:1908.10396v5 fatcat:fu73pkpl3bamhazjiyxytybwxa

Anisotropic Additive Quantization for Fast Inner Product Search

Jin Zhang, Qi Liu, Defu Lian, Zheng Liu, Le Wu, Enhong Chen
2022 AAAI Conference on Artificial Intelligence  
Maximum Inner Product Search (MIPS) plays an important role in many applications ranging from information retrieval, recommender systems to natural language processing and machine learning.  ...  The state-of-the-art approximated MIPS is product quantization with a score-aware loss, which weighs more heavily on items with larger inner product scores.  ...  Acknowledgments The work was supported by grants from the National Natural Science Foundation of China (No. 62022077 and 61976198) and the Fundamental Research Funds for the Central Universities (No.  ... 
dblp:conf/aaai/ZhangLLLWC22 fatcat:bfkwrqnobjbmnk4a3rk2vjbypu

Exemplar Loss for Siamese Network in Visual Tracking [article]

Shuo Chang, YiFan Zhang, Sai Huang, Yuanyuan Yao, Zhiyong Feng
2020 arXiv   pre-print
In this paper, a new exemplar loss integrated with logistic loss is proposed to enhance the feature model's discrimination by reducing inner products among exemplars.  ...  In general, Siamese tracking algorithms, supervised by logistic loss and triplet loss, increase the value of inner product between exemplar template and positive sample while reduce the value of inner  ...  With this new exemplar loss, the inner products among different exemplar templates are minimized.  ... 
arXiv:2006.12987v2 fatcat:3r3hwkfdqzh6her5cp5o23jcnu

Fast Matching Pursuit Video Coding by Combining Dictionary Approximation and Atom Extraction

Jian-Liang Lin, Wen-Liang Hwang, Soo-Chang Pei
2007 IEEE transactions on circuits and systems for video technology (Print)  
We combine calculation of the inner products and maximum atom extraction of a matching pursuit video coding scheme based on eigendictionary approximation and tree-based vector quantization.  ...  Index Terms-Fast algorithm, matching pursuit (MP), treebased vector quantization (VQ), video coding.  ...  ACKNOWLEDGMENT The authors would like to thank the reviewers for their valuable comments which helped us improve the presentation of the manuscript.  ... 
doi:10.1109/tcsvt.2007.903120 fatcat:ds3ndhwmbrc2pco6yyo762zmf4

Deep Visual-Semantic Quantization for Efficient Image Retrieval

Yue Cao, Mingsheng Long, Jianmin Wang, Shichen Liu
2017 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
DVSQ enables efficient and effective image retrieval by supporting maximum inner-product search, which is computed based on learned codebooks with fast distance table lookup.  ...  The main contribution lies in jointly learning deep visual-semantic embeddings and visual-semantic quantizers using carefullydesigned hybrid networks and well-specified loss functions.  ...  Big Data System Software (NEL-BDSS), and Tsinghua National Lab for Information Science and Technology (TNList) Projects.  ... 
doi:10.1109/cvpr.2017.104 dblp:conf/cvpr/CaoL0L17 fatcat:jgzhlmcoeraqblejcdeoeovh6i

Projective Clustering Product Quantization [article]

Aditya Krishnan, Edo Liberty
2021 arXiv   pre-print
This paper suggests the use of projective clustering based product quantization for improving nearest neighbor and max-inner-product vector search (MIPS) algorithms.  ...  We provide anisotropic and quantized variants of projective clustering which outperform previous clustering methods used for this problem such as ScaNN.  ...  Accurate and fast asymmetric locality-sensitive hashing scheme for maximum inner product search.  ... 
arXiv:2112.02179v1 fatcat:yxxazxoxlne7hkn5bsnvw3iwhu

Short-term Load Forecasting Model for Microgrid Based on HSA-SVM

Han Aoyang, Yu Litao, An Shuhuai, Zhang Zhisheng, J. Heled, A. Yuan
2018 MATEC Web of Conferences  
The harmony search optimization algorithm(HSA) is used to optimize the parameters of the SVM model, because it has the advantages of fast convergence speed and better optimization ability.  ...  Accurate load forecasting for microgrid will provide the necessary basis for cooperative optimization scheduling.  ...  Then, the new solution is randomly searched for each variable in the HM with the retention probability p 1 , and the new solution is searched in the feasible domain with the probability 1-p 1 .  ... 
doi:10.1051/matecconf/201817301007 fatcat:hgrivkaho5al3giby73awsfdg4

NSGA-II Algorithm with a Local Search Strategy for Multiobjective Optimal Design of Dry-Type Air-Core Reactor

Chengfen Zhang, Xikui Ma
2015 Mathematical Problems in Engineering  
Dry-type air-core reactor is now widely applied in electrical power distribution systems, for which the optimization design is a crucial issue.  ...  Furthermore, another modification is that an external archive is set in the proposed algorithm for increasing the evolutionary efficiency.  ...  In the local search procedure, let CMA denote the number of generation for local search and in and av the neighborhood size of the inner diameter and the average side length of the selected nondominated  ... 
doi:10.1155/2015/839035 fatcat:i255restk5aijbgdun5f4sd6z4

Maximum inner-product search using cone trees

Parikshit Ram, Alexander G. Gray
2012 Proceedings of the 18th ACM SIGKDD international conference on Knowledge discovery and data mining - KDD '12  
Subsequently, we present a dual-tree algorithm for the case where there are multiple queries. Our proposed branch-and-bound algorithms are based on novel inner-product bounds.  ...  Finally we present a new data structure, the cone tree, for increasing the efficiency of the dualtree algorithm.  ...  However, the final search still involves a linear scan over the set of points for the maximum inner-product or a fast nearestneighbor search under the assumption that finding the nearestneighbor is equivalent  ... 
doi:10.1145/2339530.2339677 dblp:conf/kdd/RamG12 fatcat:yzcehinodve6rjwjkbqmvgh55a

A fast fractal image coding based on kick-out and zero contrast conditions

Cheung-Ming Lai, Kin-Man Lam, Wan-Chi Siu
2003 IEEE Transactions on Image Processing  
In Section IV, we compare the performance of our proposed fast algorithm with the full search, the adaptive search scheme [12] , tree search scheme [13] , DCT Inner Product algorithm [2] , and our algorithm  ...  One example is the DCT Inner Product [2] approach, which allows the computation of two inner products only for the eight isometry operations of a domain block.  ... 
doi:10.1109/tip.2003.817246 pmid:18244697 fatcat:ayku47b3yfg7ngmyr4hlnxjsjm

On Efficient Retrieval of Top Similarity Vectors

Shulong Tan, Zhixin Zhou, Zhaozhuo Xu, Ping Li
2019 Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)  
In this paper we demonstrate an efficient method for searching vectors via a typical nonmetric matching function: inner product.  ...  Our method, which constructs an approximate Inner Product Delaunay Graph (IPDG) for top-1 Maximum Inner Product Search (MIPS), transforms retrieving the most suitable latent vectors into a graph search  ...  Acknowledgement The authors would like to sincerely thank the anonymous reviewers of NAACL 2019 and EMNLP 2019, for their helpful comments, which has improved the quality of this paper.  ... 
doi:10.18653/v1/d19-1527 dblp:conf/emnlp/TanZXL19 fatcat:hc43dkbomzdplje4asx5w7m2h4

Maximum Inner-Product Search using Tree Data-structures [article]

Parikshit Ram, Alexander G. Gray
2012 arXiv   pre-print
These branch-and-bound algorithms involve novel bounds suited for the purpose of best-matching with inner products.  ...  Subsequently, we present a dual-tree algorithm for the case where there are multiple queries. Finally we present a new data structure for increasing the efficiency of the dual-tree algorithm.  ...  Why is maximum inner-product search possibly harder?  ... 
arXiv:1202.6101v1 fatcat:lzfnkqhwozcgjk74ozzivnfhsy

Asymmetric LSH (ALSH) for Sublinear Time Maximum Inner Product Search (MIPS) [article]

Anshumali Shrivastava, Ping Li
2014 arXiv   pre-print
We present the first provably sublinear time algorithm for approximate Maximum Inner Product Search (MIPS).  ...  Our proposal is also the first hashing algorithm for searching with (un-normalized) inner product as the underlying similarity measure. Finding hashing schemes for MIPS was considered hard.  ...  Introduction and Motivation The focus of this paper is on the problem of Maximum Inner Product Search (MIPS) .  ... 
arXiv:1405.5869v1 fatcat:vh5jouptxnhg3hpmflkah34tjq

Improved Asymmetric Locality Sensitive Hashing (ALSH) for Maximum Inner Product Search (MIPS) [article]

Anshumali Shrivastava, Ping Li
2014 arXiv   pre-print
Recently it was shown that the problem of Maximum Inner Product Search (MIPS) is efficient and it admits provably sub-linear hashing algorithms.  ...  Theoretical analysis show that the new scheme is significantly better than the original scheme for MIPS. Experimental evaluations strongly support the theoretical findings.  ...  The method and theoretical analysis for Sign-ALSH were conducted right after the initial submission of our first work on ALSH [18] in February 2014.  ... 
arXiv:1410.5410v2 fatcat:h3cdbk43frbu5i5qod2l6pfo3i

Multiple blocks matching pursuit update algorithm for low bit rate video coding

Jian-Liang Lin, Wen-Liang Hwang, Soo-Chang Pei
2006 IEEE transactions on circuits and systems for video technology (Print)  
Index Terms-Content based video coding, fast algorithm, matching pursuit (MP), video coding.  ...  By approximating a residual in a subspace, we propose a multiple blocks MP update algorithm for video coding that achieves a faster and better MP approximation than traditional algorithms.  ...  In Step 2, after the atom with maximum inner product has been selected, the inner products between the new residual and the bases in must be updated.  ... 
doi:10.1109/tcsvt.2006.869969 fatcat:mh6kixrxbjc4fmfmpq4gaubp64
« Previous Showing results 1 — 15 out of 66,799 results