Filters








33,608 Hits in 5.6 sec

Fast Low-rank Metric Learning for Large-scale and High-dimensional Data [article]

Han Liu, Zhizhong Han, Yu-Shen Liu, Ming Gu
2019 arXiv   pre-print
Low-rank metric learning aims to learn better discrimination of data subject to low-rank constraints.  ...  To address this issue, we present a novel fast low-rank metric learning (FLRML) method.FLRML casts the low-rank metric learning problem into an unconstrained optimization on the Stiefel manifold, which  ...  Besides low-rank metric learning methods, there are some other types of methods for speeding up metric learning on large and high-dimensional datasets.  ... 
arXiv:1909.06297v1 fatcat:olhmi5xutbbs5kkjflnjn5tduy

A Fast Clustering Algorithm for Large-scale and High Dimensional Data

Ming LIU, Xiao-Long WANG, Yuan-Chao LIU
2009 ACTA AUTOMATICA SINICA  
Low-rank metric learning aims to learn better discrimination of data subject to low-rank constraints.  ...  To address this issue, we present a novel fast low-rank metric learning (FLRML) method.  ...  Besides low-rank metric learning methods, there are some other types of methods for speeding up metric learning on large and high-dimensional datasets.  ... 
doi:10.3724/sp.j.1004.2009.00859 fatcat:w4i6stk7drajvfbq7srh2ct2z4

Robust Discriminative Metric Learning for Image Representation

Zhengming Ding, Ming Shao, Wonjun Hwang, Sungjoov Suh, Jae-Joon Han, Changkyu Choi, Yun Fu
2019 IEEE transactions on circuits and systems for video technology (Print)  
Furthermore, fast low-rank representation is implemented to mitigate the computational burden and make sure the scalability on large-scale datasets.  ...  In this paper, we propose a Robust Discriminative Metric Learning algorithm (RDML) via fast low-rank representation and denoising strategy.  ...  However, the downside is that deep metric learning usually requires large-scale data for training.  ... 
doi:10.1109/tcsvt.2018.2879626 fatcat:o7x6pyoqcbbjbfu35dmp5uvaay

A Fast Logdet Divergence Based Metric Learning Algorithm for Large Data Sets Classification

Jiangyuan Mei, Jian Hou, Jicheng Chen, Hamid Reza Karimi
2014 Abstract and Applied Analysis  
It is a challenging task to classify large data sets efficiently, accurately, and robustly, as large data sets always contain numerous instances with high dimensional feature space.  ...  Meanwhile, we propose a compressed representation for high dimensional Mahalanobis matrix to reduce the computation complexity in each iteration.  ...  Conclusion In this paper we propose a fast and robust metric learning algorithm for large data sets classification.  ... 
doi:10.1155/2014/463981 fatcat:fl4zje4btbh73axbfmvahsnyjm

FILM: A Fast, Interpretable, and Low-rank Metric Learning Approach for Sentence Matching [article]

Xiangru Tang, Alan Aw
2020 arXiv   pre-print
To alleviate this problem, we explore a metric learning approach, named FILM (Fast, Interpretable, and Low-rank Metric learning) to efficiently find a high discriminative projection of the high-dimensional  ...  data.  ...  Because metric learning has an advantage in time and memory usage on large-scale and high-dimensional datasets compared with methods above.  ... 
arXiv:2010.05523v2 fatcat:u6h7nignerd3vba5pdzshwvnhm

Fast solvers and efficient implementations for distance metric learning

Kilian Q. Weinberger, Lawrence K. Saul
2008 Proceedings of the 25th international conference on Machine learning - ICML '08  
For large data sets, the use of locally adaptive distance metrics leads to even lower error rates.  ...  Second, we show how to reduce both training and testing times using metric ball trees; the speedups from ball trees are further magnified by learning low dimensional representations of the input space.  ...  The training procedure for LMNN-SVD optimized a full-rank distance metric in this 350 dimensional space, then extracted a low-rank distance metric from its leading eigenvectors.  ... 
doi:10.1145/1390156.1390302 dblp:conf/icml/WeinbergerS08 fatcat:w7tcorg7ovce5mjt5g34zrlkci

Fine-Grained Visual Categorization via Multi-stage Metric Learning [article]

Qi Qian, Rong Jin, Shenghuo Zhu, Yuanqing Lin
2015 arXiv   pre-print
To this end, we proposed a multi-stage metric learning framework that divides the large-scale high dimensional learning problem to a series of simple subproblems, achieving O(d) computational complexity  ...  However, feature representation of an image is often high dimensional, and DML is known to have difficulty in dealing with high dimensional feature vectors since it would require O(d^2) for storage and  ...  Acknowledgments Qi Qian and Rong Jin are supported in part by ARO (W911NF-11-1-0383), NSF (IIS-1251031) and ONR (N000141410631).  ... 
arXiv:1402.0453v2 fatcat:yrlejbqgkvgqrkjkg4oxjxbraa

Scalable Metric Learning via Weighted Approximate Rank Component Analysis [article]

Cijo Jose, Francois Fleuret
2016 arXiv   pre-print
We propose a metric learning formulation called Weighted Approximate Rank Component Analysis (WARCA).  ...  We are interested in the large-scale learning of Mahalanobis distances, with a particular focus on person re-identification.  ...  Linear spectral methods are very fast for low dimensional problems but the training time scales quadratically in the data dimension.  ... 
arXiv:1603.00370v2 fatcat:diaqkjwup5fwjk4laywxkjenqi

Fine-grained visual categorization via multi-stage metric learning

Qi Qian, Rong Jin, Shenghuo Zhu, Yuanqing Lin
2015 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)  
To this end, we proposed a multi-stage metric learning framework that divides the large-scale high dimensional learning problem to a series of simple subproblems, achieving O(d) computational complexity  ...  However, feature representation of an image is often high dimensional, and DML is known to have difficulty in dealing with high dimensional feature vectors since it would require O(d 2 ) for storage and  ...  high dimensional data, and develops a randomized low rank matrix approximation algorithm for the storage challenge.  ... 
doi:10.1109/cvpr.2015.7298995 dblp:conf/cvpr/QianJZL15 fatcat:vp5xdwp2bfcglaowe6u5kd434q

Tiny Descriptors for Image Retrieval with Unsupervised Triplet Hashing

Jie Lin, Olivier Morere, Julie Petta, Vijay Chandrasekhar, Antoine Veillard
2016 2016 Data Compression Conference (DCC)  
A good image descriptor is key to the retrieval pipeline and should reconcile two contradictory requirements: providing recall rates as high as possible and being as compact as possible for fast matching  ...  Then, triplet networks, a rank learning scheme based on weight sharing nets is used to fine-tune the binary embedding functions to retain as much as possible of the useful metric properties of the original  ...  However, SRBMs is purely based on data reconstruction and do not purposefully try to preserve the good metric properties of the original high-dimensional space.  ... 
doi:10.1109/dcc.2016.23 dblp:conf/dcc/LinMPCV16 fatcat:qxdp7hxcpbe6ffk3ck5jta45pu

Tiny Descriptors for Image Retrieval with Unsupervised Triplet Hashing [article]

Jie Lin, Olivier Morère, Julie Petta, Vijay Chandrasekhar, Antoine Veillard
2015 arXiv   pre-print
A good image descriptor is key to the retrieval pipeline and should reconcile two contradictory requirements: providing recall rates as high as possible and being as compact as possible for fast matching  ...  Following the recent successes of Deep Convolutional Neural Networks (DCNN) for large scale image classification, descriptors extracted from DCNNs are increasingly used in place of the traditional hand  ...  However, SRBMs is purely based on data reconstruction and do not purposefully try to preserve the good metric properties of the original high-dimensional space.  ... 
arXiv:1511.03055v1 fatcat:owk7tvr3ibectc6f5u2knokggi

An empirical evaluation of supervised learning in high dimensions

Rich Caruana, Nikos Karampatziakis, Ainur Yessenalina
2008 Proceedings of the 25th international conference on Machine learning - ICML '08  
We evaluate performance on three metrics: accuracy, AUC, and squared loss and study the effect of increasing dimensionality on the performance of the learning algorithms.  ...  Our findings are consistent with previous studies for problems of relatively low dimension, but suggest that as dimensionality increases the relative performance of the learning algorithms changes.  ...  We also thank Alec Berntson Eric Breck and Art Munson for providing the crystallography, DSE and ornithology datasets respectively. Art also put together the calibration procedures.  ... 
doi:10.1145/1390156.1390169 dblp:conf/icml/CaruanaKY08 fatcat:cqudewtkmnevvb3vravymdpbla

Cholesky Decomposition Based Metric Learning for Video-based Human Action Recognition

Si Chen, Yuanyuan Shen, Yan Yan, Dahan Wang, Shunzhi Zhu
2020 IEEE Access  
Different from the traditional low-rank metric learning methods that explicitly adopt the low-rank constraint to learn the matrix, the proposed algorithm achieves such a constraint by controlling the rank  ...  Metric learning, which learns a similarity metric, plays an important role in human action recognition. However, learning a full-rank matrix is usually inefficient and easily leads to overfitting.  ...  Although the excellent performance achieved by the deep learning based methods, these methods often require the large-scale training data, and thus the training complexity is relatively high [11] .  ... 
doi:10.1109/access.2020.2966329 fatcat:vkekokypynghhfohktkeihs5y4

Comprehensive study and Analysis of Extreme Multi-Label Classification Approach

Purvi Prajapati
2020 International Journal of Advanced Trends in Computer Science and Engineering  
This paper discussed different approaches for large scale Recommendation System using Extreme Multi-Label Classification Approach and empirical evaluation carried out on three multi-label datasets which  ...  handles large volume of the data.  ...  In embedding based approach reduces data from high dimensionality into low dimensionality. This approach has exploit label co-relations and data sparsity to compress the number of labels.  ... 
doi:10.30534/ijatcse/2020/83922020 fatcat:qzgi7mtk4bfnbk4dmdjblp373y

Efficient and Scalable Information Geometry Metric Learning

Wei Wang, Bao-Gang Hu, Zengfu Wang
2013 2013 IEEE 13th International Conference on Data Mining  
Following this property, SIGML is found to be capable of handling both full-rank and low-rank kernels. Additionally, the geometric information from data is further exploited in SIGML.  ...  Information Geometry Metric Learning (IGML) is shown to be an effective algorithm for distance metric learning.  ...  The running time increases rapidly for high-dimensional data. (2) IGML employs a (potential) low-rank ideal kernel [13] .  ... 
doi:10.1109/icdm.2013.67 dblp:conf/icdm/WangHW13 fatcat:nwx4loaomnclfcgp7j33cprwea
« Previous Showing results 1 — 15 out of 33,608 results