检索结果-内蒙古大学图书馆

iSplit LBI: Individualized partial ranking with ties via split LBI

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Xu, Qianqian Sun, Xinwei Yang, Zhiyong Cao, Xiaochun Huang, Qingming Yao, Yuan Key Lab. of Intelligent Information Processing Institute of Computing Technology CAS Microsoft Research Asia State Key Laboratory of Information Security Institute of Information Engineering CAS School of Cyber Security University of Chinese Academy of Sciences School of Computer Science and Tech. University of Chinese Academy of Sciences Key Laboratory of Big Data Mining and Knowledge Management CAS Peng Cheng Laboratory Department of Mathematics Hong Kong University of Science and Technology Hong Kong

Due to the inherent uncertainty of data, the problem of predicting partial ranking from pairwise comparison data with ties has attracted increasing interest in recent years. However, in real-world scenarios, different individuals often hold distinct preferences. It might be misleading to merely look at a global partial ranking while ignoring personal diversity. In this paper, instead of learning a global ranking which is agreed with the consensus, we pursue the tie-aware partial ranking from an individualized perspective. Particularly, we formulate a unified framework which not only can be used for individualized partial ranking prediction, but also be helpful for abnormal user selection. This is realized by a variable splitting-based algorithm called iSplitLBI. Specifically, our algorithm generates a sequence of estimations with a regularization path, where both the hyperparameters and model parameters are updated. At each step of the path, the parameters can be decomposed into three orthogonal parts, namely, abnormal signals, personalized signals and random noise. The abnormal signals can serve the purpose of abnormal user selection, while the abnormal signals and personalized signals together are mainly responsible for individual partial ranking prediction. Extensive experiments on simulated and real-world datasets demonstrate that our new approach significantly outperforms state-of-the-art alternatives. The code is now availiable at https://***/qianqianxu010/ NeurIPS2019-iSplitLBI. Copyright © 2019, The Authors. All rights reserved.

关键词： Forecasting

Larger, Cheaper, but Faster: SSD-SMR hybrid storage boosted by a new SMR-oriented cache framework 33

学校读者我要写书评

暂无评论

Larger, Cheaper, but Faster: SSD-SMR hybrid storage boosted ...

33rd International Conference on Massive Storage Systems and Technology, MSST 2017

作者： Wang, Chunling Wang, Dandan Chai, Yupeng Wang, Chuanwen Sun, Diansen Key Laboratory of Data Engineering and Knowledge Engineering School of Information Renmin University of China Beijing China

By utilizing the new Shingled Magnetic Recording (SMR) technique, the emerging SMR disks achieve higher storage density and lower costs. Together with Flash-based SSDs, SMR disks can be used to construct a new hybrid storage system that is larger, but cheaper than the widely used conventional diskbased storage systems. It is appropriate for the fast growing big data applications which have strict requirements on both capacity and cost, e.g., the scientific computing in the fields of meteorology and astronomy. However, the most serious challenge of the SSD-SMR hybrid storage lies in its weak fine-grained random write performance caused by the write amplification in SMR disks, especially when I/Os are spread randomly across the Logical Block Address (LBA) range of the entire disk. Traditional cache algorithms like LRU usually lead to tens of or even hundreds of times more data written into magnetic plates than the original requests, thus resulting in poor I/O performance, because they do not consider and limit the LBA distribution of the eviction victims at all. Therefore, the cache algorithms for SSD-SMR hybrid storage systems should take this new dimension into account except for the data popularity. In this paper, we propose a new SMR-oriented cache framework, i.e., Partially Open Region for Eviction (PORE), which restricts the LBA range of evicted data from SSD caches appropriately to promote the overall performance of the hybrid storage by making a better tradeoff between the cached data popularity and the SMR write amplification. The experiments based on real-world traces reveal that PORE can boost the SSD-SMR hybrid storage performance by a factor of 3.26∼10.2 with only a 2% SSD cache compared with a conventional disk-based storage on average, while the hybrid storage coupled with LRU is slower than the conventional storage sometimes. PORE makes the hybrid storage much faster than traditional storage besides the advantages on capacity and cost. © 33rd Internati

关键词： Hard disk storage

Adaptive structure-constrained robust latent low-rank coding for image recovery

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Zhang, Zhao Wang, Lei Li, Sheng Wang, Yang Zhang, Zheng Zha, Zhengjun Wang, Meng School of Computer Science and Technology Soochow University Suzhou215006 China Key Laboratory of Knowledge Engineering with Big Data Ministry of Education Hefei University of Technology School of Computer Science and Information Engineering Hefei University of Technology Hefei China Department of Computer Science University of Georgia 549 Boyd GSRC AthensGA30602 Shenzhen China School of Information Science and Technology University of Science and Technology of China Hefei China

In this paper, we propose a robust representation learning model called Adaptive Structure-constrained Low-Rank Coding (AS-LRC) for the latent representation of data. To recover the underlying subspaces more accurately, AS-LRC seamlessly integrates an adaptive weighting based block-diagonal structure-constrained low-rank representation and the group sparse salient feature extraction into a unified framework. Specifically, AS-LRC performs the latent decomposition of given data into a low-rank reconstruction by a block-diagonal codes matrix, a group sparse locality-adaptive salient feature part and a sparse error part. To enforce the block-diagonal structures adaptive to different real datasets for the low-rank recovery, AS-LRC clearly computes an auto-weighting matrix based on the locality-adaptive features and multiplies by the low-rank coefficients for direct minimization at the same time. This encourages the codes to be block-diagonal and can avoid the tricky issue of choosing optimal neighborhood size or kernel width for the weight assignment, suffered in most local geometrical structures-preserving low-rank coding methods. In addition, our AS-LRC selects the L2,1-norm on the projection for extracting group sparse features rather than learning low-rank features by Nuclear-norm regularization, which can make learnt features robust to noise and outliers in samples, and can also make the feature coding process efficient. Extensive visualizations and numerical results demonstrate the effectiveness of our AS-LRC for image representation and recovery. Copyright © 2019, The Authors. All rights reserved.

关键词： Recovery

Notes on spatial twisted central configurations for 2N-body problem

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Ding, Liang Sánchez-Cerritos, Juan Manuel Wei, Jinlong School of Data Science and Information Engineering Guizhou Minzu University Guiyang550025 College of Mathematics and Statistics Chongqing Technology and Business University Chongqing400067 Chongqing Key Laboratory of Social Economy and Applied Statistics Chongqing400067 School of Statistics and Mathematics Zhongnan University of Economics and Law Wuhan430073

We study the spatial central configuration formed by two twisted regular N-polygons with any twist angle θ, and prove that the sizes of the two regular N-polygons must be equal to each *** Codes 70F07, 70F15 Copyrigh...

关键词：

Fully-convolutional intensive feature flow neural network for text recognition

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Zhang, Zhao Tang, Zemin Zhang, Zheng Wang, Yang Qin, Jie Wang, Meng School of Computer Science and Technology Soochow University China Key Laboratory of Knowledge Engineering with Big Data Ministry of Education School of Computer and Information Hefei University of Technology Hefei China Bio-Computing Research Center Harbin Institute of Technology Shenzhen518055 China Inception Institute of Artificial Intelligence Abu Dhabi United Arab Emirates

The Deep Convolutional Neural Networks (CNNs) have obtained a great success for pattern recognition, such as recognizing the texts in images. But existing CNNs based frameworks still have several drawbacks: 1) the traditaional pooling operation may lose important feature information and is unlearnable;2) the traditional convolution operation optimizes slowly and the hierarchical features from different layers are not fully utilized. In this work, we address these problems by developing a novel deep network model called Fully-Convolutional Intensive Feature Flow Neural Network (IntensiveNet). Specifically, we design a further dense block called intensive block to extract the feature information, where the original inputs and two dense blocks are connected tightly. To encode data appropriately, we present the concepts of dense fusion block and further dense fusion operations for our new intensive block. By adding short connections to different layers, the feature flow and coupling between layers are enhanced. We also replace the traditional convolution by depthwise separable convolution to make the operation efficient. To prevent important feature information being lost to a certain extent, we use a convolution operation with stride 2 to replace the original pooling operation in the customary transition layers. The recognition results on large-scale Chinese string and MNIST datasets show that our IntensiveNet can deliver enhanced recognition results, compared with other related deep models. Copyright © 2019, The Authors. All rights reserved.

关键词： Convolution

Adaptive Structure-Constrained Robust Latent Low-Rank Coding for Image Recovery

学校读者我要写书评

暂无评论

Adaptive Structure-Constrained Robust Latent Low-Rank Coding...

IEEE International Conference on data Mining (ICDM)

作者： Zhao Zhang Lei Wang Sheng Li Yang Wang Zheng Zhang Zhengjun Zha Meng Wang School of Computer Science and Technology Soochow University Suzhou China Key Laboratory of Knowledge Engineering with Big Data (Ministry of Education) Hefei University of Technology School of Computer Science and Information Engineering Hefei University of Technology Hefei China Department of Computer Science University of Georgia Athens GA Bio-Computing Research Center Harbin Institute of Technology (Shenzhen) Shenzhen China School of Information Science and Technology University of Science and Technology of China Hefei China

In this paper, we propose a robust representation learning model called Adaptive Structure-constrained Low-Rank Coding (AS-LRC) for the latent representation of data. To recover the underlying subspaces more accurately, AS-LRC seamlessly integrates an adaptive weighting based block-diagonal structure-constrained low-rank representation and the group sparse salient feature extraction into a unified framework. Specifically, AS-LRC performs the latent decomposition of given data into a low-rank reconstruction by a block-diagonal codes matrix, a group sparse locality-adaptive salient feature part and a sparse error part. To enforce the block-diagonal structures adaptive to different real datasets for the low-rank recovery, AS-LRC clearly computes an auto-weighting matrix based on the locality-adaptive features and multiplies by the low-rank coefficients for direct minimization at the same time. This encourages the codes to be block-diagonal and can avoid the tricky issue of choosing optimal neighborhood size or kernel width for the weight assignment, suffered in most local geometrical structures-preserving low-rank coding methods. In addition, our AS-LRC selects the L2, 1-norm on the projection for extracting group sparse features rather than learning low-rank features by Nuclear-norm regularization, which can make learnt features robust to noise and outliers in samples, and can also make the feature coding process efficient. Extensive visualizations and numerical results demonstrate the effectiveness of our AS-LRC for image representation and recovery.

关键词：

A Novel Texture Generation Super Resolution Model

学校读者我要写书评

暂无评论

Procedia Computer Science 2019年 162卷 924-931页

作者： Biao Li Yong Shi Sujuan Li Bo Wang Zhiquan Qi Jiabin Liu School of Economics and Management University of Chinese Academy of Sciences Beijing 101408 China Research Center on Fictitious Economy and Data Science Chinese Academy of Sciences Beijing 100190 China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing 100190 China College of Information Science and Technology University of Nebraska at Omaha NE 68182 USA Department of Otology the First Affiliated Hospital of Zhengzhou University No. 1 eastern Jianshe road Zhengzhou Henan 450052 China School of Information Technology and Management University of International Business and Economics Beijing 100029 China Department of Computer Science and Engineering Texas A&M University TX 77843 USA

Recently, super-resolution methods pursue visual pleasant details attract more attention in academic circle. Unlike former accurate driving models, they leverage new losses measured the difference of features extracting from a pre-trained deep learning model. Moreover, the popular GANs structure are introduced to approach better outputs. In general, these new visual feeling promoting losses are constrained by the learning ability of model and effected by the weakness of GANs. In order to address these problems, we propose a novel SR method to generating abundant details. Based on former SR methods and texture generated models, we combine new model constructure and integrated loss function to achieve our goal. The new constructure derives from the very thought that more information from the LR input leads to more accurate results. The new integrated loss try to decrease the weakness of each one by corresponding loss.. Experiments show that our model achieves higher quality generations than former methods.

关键词： Super Resolution Multi Scale Smooth Loss Visual Quality

NaCdSb: An Orthorhombic Zintl Phase with Exceptional Intrinsic Thermoelectric Performance

学校读者我要写书评

暂无评论

Angewandte Chemie 2022年第3期135卷

作者： Kai Guo Yuting Zhang Song Yuan Qinghang Tang Chen Lin Pengfei Luo Jiong Yang Shusheng Pan Li-Dong Zhao Guofeng Cheng Jiye Zhang Jun Luo School of Physics and Materials Science Guangzhou University Guangzhou 510006 China Research Center for Advanced Information Materials (CAIM) Huangpu Research & Graduate School of Guangzhou University Sino-Singapore Guangzhou Knowledge City Huangpu District Guangzhou 510555 China Contribution: Conceptualization (lead) Formal analysis (equal) Writing - original draft (lead) School of Materials Science and Engineering Shanghai University 99 Shangda Road Shanghai 200444 China Contribution: Data curation (equal) Investigation (equal) Validation (lead) Contribution: Data curation (lead) Formal analysis (equal) Materials Genome Institute Shanghai University 99 Shangda Road Shanghai 200444 China Key Laboratory of Artificial Micro- and Nano-Structures of Ministry of Education and School of Physics and Technology Wuhan University Wuhan 430072 China Contribution: Investigation (equal) Contribution: Investigation (equal) Methodology (supporting) Contribution: Investigation (equal) Writing - review & editing (supporting) School of Materials Science and Engineering Beihang University Beijing 100191 China Contribution: Methodology (equal) Writing - review & editing (equal) Analysis & Testing Center for Inorganic Materials Shanghai Institute of Ceramics Chinese Academy of Sciences Shanghai 200050 China Contribution: Formal analysis (equal) Investigation (supporting) Contribution: Data curation (equal) Writing - review & editing (supporting) Contribution: Supervision (lead) Writing - review & editing (equal)

Many Zintl phases are promising thermoelectric materials owning to their features like narrow band gaps, multiband behaviors, ideal charge transport tunnels, and loosely bound cations. Herein we show a new Zintl phase NaCdSb with exceptional intrinsic thermoelectric performance. Pristine NaCdSb exhibits semiconductor behaviors with an experimental hole concentration of 2.9×10 18 cm −3 and a calculated band gap of 0.5 eV. As the temperature increases, the hole concentration rises gradually and approaches its optimal one, leading to a high power factor of 11.56 μW cm −1 K −2 at 673 K. The ultralow thermal conductivity is derived from the small phonon group velocity and short phonon lifetime, ascribed to the structural anharmonicity of Cd−Sb bonds. As a consequence, a maximum zT of 1.3 at 673 K has been achieved without any doping optimization or structural modification, demonstrating that NaCdSb is a remarkable thermoelectric compound with great potential for performance improvement.

关键词： Electron-Crystal NaCdSb Phonon-Glass Thermoelectric Properties Zintl Phase

s-LWSR: Super lightweight super-resolution network

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Li, Biao Liu, Jiabin Wang, Bo Qi, Zhiquan Shi, Yong School of Economics and Management University of Chinese Academy of Sciences Beijing101408 China Research Center on Fictitious Economy and Data Science Chinese Academy of Sciences Beijing100190 China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Beijing100190 China School of Information Technology and Management University of International Business and Economics Beijing100029 China Department of Computer Science and Engineering Texas A&M University College StationTX77843 United States College of Information Science and Technology University of Nebraska OmahaNE68182 United States

Deep learning (DL) architectures for super-resolution (SR) normally contain tremendous parameters, which has been regarded as the crucial advantage for obtaining satisfying performance. However, with the widespread use of mobile phones for taking and retouching photos, this character greatly hampers the deployment of DL-SR models on the mobile devices. To address this problem, in this paper, we propose a super lightweight SR network: s-LWSR. There are mainly three contributions in our work. Firstly, in order to efficiently abstract features from the low resolution image, we build an information pool to mix multi-level information from the first half part of the pipeline. Accordingly, the information pool feeds the second half part with the combination of hierarchical features from the previous layers. Secondly, we employ a compression module to further decrease the size of parameters. Intensive analysis confirms its capacity of trade-off between model complexity and accuracy. Thirdly, by revealing the specific role of activation in deep models, we remove several activation layers in our SR model to retain more information for performance improvement. Extensive experiments show that our s-LWSR, with limited parameters and operations, can achieve similar performance to other cumbersome DL-SR methods. Copyright © 2019, The Authors. All rights reserved.

关键词： Chemical activation