Filters








10,151 Hits in 6.1 sec

Scalable and Efficient Pairwise Learning to Achieve Statistical Accuracy

Bin Gu, Zhouyuan Huo, Heng Huang
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Existing pairwise learning algorithms do not perform well in the generality, scalability and efficiency simultaneously.  ...  To address these challenging problems, in this paper, we first analyze the relationship between the statistical accuracy and the regularized empire risk for pairwise loss.  ...  The scalability and efficiency are still the bottlenecks of existing pairwise learning algorithms.  ... 
doi:10.1609/aaai.v33i01.33013697 fatcat:c3zaqqwmwjgr5dx4gfqu7z6pp4

WMRB: Learning to Rank in a Scalable Batch Training Approach [article]

Kuan Liu, Prem Natarajan
2017 arXiv   pre-print
We propose a new learning to rank algorithm, named Weighted Margin-Rank Batch loss (WMRB), to extend the popular Weighted Approximate-Rank Pairwise loss (WARP).  ...  WMRB uses a new rank estimator and an efficient batch training algorithm.  ...  However, it is not scalable to large item set in practice due to its intrinsic online learning fashion.  ... 
arXiv:1711.04015v1 fatcat:jrmayomu6vetbi5pfuvmc73ccy

On Training Knowledge Graph Embedding Models

Sameh K. Mohamed, Emir Muñoz, Vit Novacek
2021 Information  
., RDF graphs) in an efficient and scalable manner. The key to success of these models is their ability to learn low-rank vector representations for knowledge graph entities and relations.  ...  We finally investigate the effects of specific choices on the scalability and accuracy of knowledge graph embedding models.  ...  This approach provides scalable and efficient embeddings learning as it has linear time and space complexity.  ... 
doi:10.3390/info12040147 fatcat:dsavhnvr5zcuflnu3c5dp5ap5y

SimpleNPKL

Jinfeng Zhuang, Ivor W. Tsang, Steven C. H. Hoi
2009 Proceedings of the 26th Annual International Conference on Machine Learning - ICML '09  
In contrast to the previous approaches, our empirical results show that our new technique achieves the same accuracy, but is significantly more efficient and scalable.  ...  In this paper, we propose an efficient approach to NPK learning from side information, referred to as SimpleNPKL, which can efficiently learn non-parametric kernels from large sets of pairwise constraints  ...  Acknowledgments This research was in part supported by Singapore MOE AcRF Tier-1 Research Grant (RG15/08) and Research Grant (RG67/07).  ... 
doi:10.1145/1553374.1553537 dblp:conf/icml/ZhuangTH09 fatcat:2g7sdd7qwfgldhcpcnyb5r6al4

Private Two-Party Cluster Analysis Made Formal Scalable [article]

Xianrui Meng, Dimitrios Papadopoulos, Alina Oprea, Nikos Triandopoulos
2019 arXiv   pre-print
Crucially, our solution employs modular design and judicious use of cryptography to achieve high degrees of efficiency and extensibility.  ...  For example, end-to-end execution of our secure approximate protocol, over 1M 10-dimensional records, completes in 35 sec, transferring only 896KB and achieving 97.09% accuracy.  ...  We combine our protocols with efficient approximate clustering in order to achieve the best of both worlds: strong security guarantees and scalability.  ... 
arXiv:1904.04475v2 fatcat:mss4mujjgngbheypvv6rurb7im

Scalable Multi-grained Cross-modal Similarity Query with Interpretability

Mingdong Zhu, Derong Shen, Lixin Xu, Xianfang Wang
2021 Data Science and Engineering  
queries. (4) A distributed query algorithm is proposed to improve the scalability of our framework.  ...  Existing researches generally focus on query accuracy by designing complex deep neural network models and hardly consider query efficiency and interpretability simultaneously, which are vital properties  ...  Acknowledgements We would like to thank selfless friends and professional reviewers for all the insightful advices.  ... 
doi:10.1007/s41019-021-00162-4 fatcat:7tdgbtoq2jc45ixrdltrl4nofu

TOD: GPU-accelerated Outlier Detection via Tensor Operations [article]

Yue Zhao, George H. Chen, Zhihao Jia
2022 arXiv   pre-print
We propose TOD, a system for efficient and scalable outlier detection (OD) on distributed multi-GPU machines.  ...  This decomposition enables TOD to accelerate OD computations by leveraging recent advances in deep learning infrastructure in both hardware and software.  ...  Zhihao Jia is partially supported by an National Science Foundation award CNS-2147909, and a Tang Endowment.  ... 
arXiv:2110.14007v2 fatcat:5fwqcku3z5ettdus3hlwu5wft4

Scalable Learning of Non-Decomposable Objectives [article]

Elad ET. Eban, Mariano Schain, Alan Mackey, Ariel Gordon, Rif A. Saurous, Gal Elidan
2017 arXiv   pre-print
Modern retrieval systems are often driven by an underlying machine learning model. The goal of such systems is to identify and possibly rank the few most relevant items for a given query or context.  ...  In practice, due to the scalability limitations of existing approaches for optimizing such objectives, large-scale retrieval systems are instead trained to maximize classification accuracy, in the hope  ...  Finally, and most importantly, our bounds give rise to an optimization approach for non-decomposable learning metrics that is highly scalable and that is applicable to truly large datasets.  ... 
arXiv:1608.04802v2 fatcat:v5b2le5e3bbzxk3fqavqncxipe

On the Accuracy and Scalability of Probabilistic Data Linkage Over the Brazilian 114 Million Cohort

Robespierre Pita, Clicia Pinto, Samila Sena, Rosemeire Fiaccone, Leila Amorim, Sandra Reis, Mauricio L. Barreto, Spiros Denaxas, Marcos Ennes Barreto
2018 IEEE journal of biomedical and health informatics  
In this paper, we present AtyImo, a hybrid probabilistic linkage tool optimized for high accuracy and scalability in massive data sets.  ...  In terms of scalability, we present AtyImo's ability to link the entire cohort in less than nine days using Spark and scaling up to 20 million records in less than 12s over heterogeneous (CPU+GPU) architectures  ...  We discuss and evaluate accuracy, scalability and performance results achieved in experimental and real scenarios.  ... 
doi:10.1109/jbhi.2018.2796941 pmid:29505402 pmcid:PMC7198121 fatcat:cuvg5hzwqvgfpnlz77on64asnq

Learning Sparse Log-Ratios for High-Throughput Sequencing Data [article]

Elliott Gordon-Rodriguez, Thomas P Quinn, John P Cunningham
2021 bioRxiv   pre-print
However, the space of these log-ratios grows combinatorially with the dimension of the input, and as a result, existing learning algorithms do not scale to increasingly common high-dimensional datasets  ...  As well as dramatically reducing runtime, our method outperforms its competitors in terms of sparsity and predictive accuracy, as measured across a wide range of benchmark datasets.  ...  SCALABILITY INTERPRETABILITY SPARSITY ACCURACY CODACORE (OURS) + + + + PAIRWISE LOG-RATIOS (GREENACRE, 2019B) − + + − SELBAL (RIVERA-PINTO ET AL., 2018) Table 2 . 2 Evaluation metrics shown for each  ... 
doi:10.1101/2021.02.11.430695 fatcat:hepgni7uabbpnl6so32j5hybfq

GRAIL

John Paparrizos, Michael J. Franklin
2019 Proceedings of the VLDB Endowment  
To address this major drawback, we present GRAIL, a generic framework to learn compact time-series representations that preserve the properties of a user-specified comparison function.  ...  The effectiveness and the scalability of time-series mining techniques critically depend on design choices for three components responsible for (i) representing; (ii) comparing; and (iii) indexing time  ...  We also thank Christos Faloutsos and Eamonn Keogh for useful discussions and Luis Gravano and Daniel Hsu for invaluable feedback.  ... 
doi:10.14778/3342263.3342648 fatcat:m7gtrlakgzb4peibo2wvtszvq4

Detection of stealthy malware activities with traffic causality and scalable triggering relation discovery

Hao Zhang, Danfeng Daphne Yao, Naren Ramakrishnan
2014 Proceedings of the 9th ACM symposium on Information, computer and communications security - ASIA CCS '14  
We use these triggering relations to reason the occurrences of network events and to pinpoint stealthy malware activities. We define a new problem of triggering relation discovery of network events.  ...  Our solution is based on domain-knowledge guided advanced learning algorithms.  ...  Our approach utilizes probabilistic machine learning algorithms and achieves high scalability and detection accuracy. We introduce a scalable feature extraction method referred to as Pairing.  ... 
doi:10.1145/2590296.2590309 dblp:conf/ccs/ZhangYR14 fatcat:kffyvrimzrca5laimnsovbcbv4

PQ-WGLOH: A bit-rate scalable local feature descriptor

Chunyu Wang, Ling-Yu Duan, Yizhou Wang, Wen Gao
2012 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
The descriptor achieves about 95% bits reduction compared with 128-Byte SIFT and allows adaptation of descriptor lengths to support user required performance.  ...  We achieve competing matching and retrieval performance with SIFT, GLOH with much fewer bits.  ...  The overlap ratio for PQ-WGLOH is 0.8604, which is slightly worse than SIFT and GLOH, with 0.8902 and 0.8729 respectively. CHoG achieves localization accuracy of 0.8134.  ... 
doi:10.1109/icassp.2012.6288040 dblp:conf/icassp/WangDWG12 fatcat:6eiylmgobjhpbckkuqltuprhpy

Robust continuous clustering

Sohil Atul Shah, Vladlen Koltun
2017 Proceedings of the National Academy of Sciences of the United States of America  
We present a clustering algorithm that achieves high accuracy across multiple domains and scales efficiently to high dimensions and large datasets.  ...  Our method achieves high accuracy across all datasets, outperforming the best prior algorithm by a factor of 3 in average rank. clustering | data analysis | unsupervised learning  ...  that can untangle mixed clusters, and optimization is performed by efficient and scalable numerical methods.  ... 
doi:10.1073/pnas.1700770114 pmid:28851838 pmcid:PMC5603997 fatcat:ytm2vunu4nfl3cx3tblsqgifdy

Spatio-temporal patterns in network events

Ting Wang, Mudhakar Srivatsa, Dakshi Agrawal, Ling Liu
2010 Proceedings of the 6th International COnference on - Co-NEXT '10  
Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation here on.  ...  The first author and the last author are partially supported by grants from NSF CyberTrust program, NSF NetSE program, an IBM SUR grant, and a grant from Intel research council.  ...  On both synthetic and real event datasets, Tar achieves an accuracy above 0.75, under missing 40% percent of symptom events.  ... 
doi:10.1145/1921168.1921172 dblp:conf/conext/WangSAL10 fatcat:mruamubuebb3hbfp3s3ocrmunq
« Previous Showing results 1 — 15 out of 10,151 results