检索结果-内蒙古大学图书馆

DiSAN: Directional self-attention network for RNN/CNN-free language understanding

学校读者我要写书评

暂无评论

arXiv 2017年

作者： Shen, Tao Jiang, Jing Zhou, Tianyi Pan, Shirui Long, Guodong Zhang, Chengqi Centre of Artificial Intelligence Feit University of Technology Sydney Paul G. Allen School of Computer Science and Engineering University of Washington

Recurrent neural nets (RNN) and convolutional neural nets (CNN) are widely used on NLP tasks to capture the long-term and local dependencies, respectively. Attention mechanisms have recently attracted enormous interest due to their highly parallelizable computation, significantly less training time, and flexibility in modeling *** propose a novel attention mechanism in which the attention between elements from input sequence(s) is directional and multi-dimensional (i.e., feature-wise). A light-weight neural net, "Directional Self-Attention Network (DiSAN)", is then proposed to learn sentence embedding, based solely on the proposed attention without any RNN/CNN structure. DiSAN is only composed of a directional self-attention with temporal order encoded, followed by a multi-dimensional attention that compresses the sequence into a vector representation. Despite its simple form, DiSAN outperforms complicated RNN models on both prediction quality and time efficiency. It achieves the best test accuracy among all sentence encoding methods and improves the most recent best result by 1:02% on the Stanford Natural Language Inference (SNLI) dataset, and shows stateof- the-art test accuracy on the Stanford Sentiment Treebank (SST), Multi-Genre natural language inference (MultiNLI), Sentences Involving Compositional Knowledge (SICK), Customer Review, MPQA, TREC question-type classification and Subjectivity (SUBJ) datasets. Copyright © 2017, The Authors. All rights reserved.

关键词： Recurrent neural networks

ICMLCA 2021 - 2nd International Conference on Machine Learni...

学校读者我要写书评

暂无评论

Preface

ICMLCA 2021 - 2nd International Conference on Machine Learning and computer Application 2021年 iii页

作者： Ning, Xiansheng Feng, Yongxin Zhang, Wenbo Yue, Mingkai Du, Ke-Lin Huang, Shutao Zhou, Fang Ba, Shuhong Fu, Chong Shui, Penglang Xu, Shuwen Wu, Sheng Wang, Shushan Du, Zhonghua Gao, Zhijun Fan, Chunlong Li, Yufeng Wang, Yupeng Liu, Qingli Yu, Jingang Yin, Zhenyu Zhang, Deyu Gao, Hongwei Wen, Feng Guo, Cean Hao, Yongping Chan, Yung-Kuan Pavlovich, Vinogradov Gennady Kirpichnikova, Irina Qian, Zhihong Ivanov, Dmitriy V. Zhou, Xueguan Nezamabadi-Pour, Hossein Yang, Qing Zhang, Yongtang Liao, Haibin Huang, Ying Srinivasan, Kathiravan Xu, Zhenghua Pachori, Ram Bilas Zhang, Jialu AlShawabkeh, Mahmoud Guo, Chengjun Chaugule, Archana Ajit Lima, Alberto Sampaio Liu, Huayong Vidal, Jorge Maestre Alrshah, Mohamed A. Ghafoor, Kayhan Zrar Kubicek, Jan Chaugule, Archana Ajit Rodríguez-Pérez, Miguel Kumar, Gulshan Nasri, Mehdi Borrego, Carlos Persico, Valerio Wagh, Sanjeev Yan, Wenjun Yang, Li Yang, Dawei Xing, Yongkang Yang, Huadong Shang, Chengguo Chang, Xingong Shi, Hongbo Li, Aijun Feng, Liping Lv, Yali Liang, Minfu Wang, Kaixuan Bai, Zengliang Liang, Anhui Nan, Zhihong Ji, Suqin Chang, Liwei Zhou, Hongbin Yao, Chunli Cao, Guodong Zhang, Juling School of Information Science and Engineering Shenyang Ligong University China School of Equipment Engineering Shenyang Ligong University China Concordia University Canada School of Automotive and Transportation Shenyang Ligong University China School of Information Science and Engineering Shenyang Ligong University China School of Equipment Engineering Shenyang Ligong University China Northeastern University China Xidian University China Beijing University of Posts and Telecommunications China Beijing Institute of Technology China Nanjing University of Science & Technology China Shenyang Jianzhu University China Shenyang Aerospace University China Dalian University China Shenyang Institute of Computing Technology Chinese Academy of Sciences China Shenyang Ligong University China National Chun Hsing University Taiwan Department of Informatics and Applied Mathematics Russia South Ural State University Russia Jilin University China Samara State University of Transport Russia Naval Universit of Engineering China Shahid Bahonar University of Kermanto Iran Central China Normal University China Guangdong Neusoft Institute China Hubei University of Science and Technology China Liuzhou Railway Vocational Technical College China Vellore Institute of Technology India Hebei University of Technology China Indian Institute of Technology Indore India Xiangnan University China Guangxi Normal University for Nationalities China UESTC China Pimpri-Chinchwad College of Engineering & Research India Federal University of Ceará Brazil Complutense University of Madrid Spain Universiti Putra Malaysia Malaysia Shanghai Jiao Tong University China VŠB - Technical University of Ostrava Cuw Pimpri-Chinchwad College of Engineering & Research Ravet Pune India University of Vigo Spain Division of Research and Development India Islamic Azad University Iran Departament Enginyeria of Information Comunicacions Autonomous University of Barcelona Spain Opportunistic Networks United States Univers

Measurement of the multiplicity dependence of Υ production ratios in pp collisions at = 13 TeV

学校读者我要写书评

暂无评论

Journal of High Energy Physics 2025年第5期2025卷 1-24页

作者： Aaij, R. Abdelmotteleb, A. S. W. Abellan Beteta, C. Abudinén, F. Ackernley, T. Adefisoye, A. A. Adeva, B. Adinolfi, M. Adlarson, P. Agapopoulou, C. Aidala, C. A. Ajaltouni, Z. Akar, S. Akiba, K. Albicocco, P. Albrecht, J. Alessio, F. Alexander, M. Aliouche, Z. Alvarez Cartelle, P. Amalric, R. Amato, S. Amey, J. L. Amhis, Y. An, L. Anderlini, L. Andersson, M. Andreianov, A. Andreola, P. Andreotti, M. Andreou, D. Anelli, A. Ao, D. Archilli, F. Argenton, M. Arguedas Cuendis, S. Artamonov, A. Artuso, M. Aslanides, E. Ataíde Da Silva, R. Atzeni, M. Audurier, B. Bacher, D. Bachiller Perea, I. Bachmann, S. Bachmayer, M. Back, J. J. Baladron Rodriguez, P. Balagura, V. Balboni, A. Baldini, W. Balzani, L. Bao, H. Baptista de Souza Leite, J. Barbero Pretel, C. Barbetti, M. Barbosa, I. R. Barlow, R. J. Barnyakov, M. Barsuk, S. Barter, W. Bartolini, M. Bartz, J. Basels, J. M. Bashir, S. Bassi, G. Batsukh, B. Battista, P. B. Bay, A. Beck, A. Becker, M. Bedeschi, F. Bediaga, I. B. Behling, N. A. Belin, S. Belous, K. Belov, I. Belyaev, I. Benane, G. Bencivenni, G. Ben-Haim, E. Berezhnoy, A. Bernet, R. Bernet Andres, S. Bertolin, A. Betancourt, C. Betti, F. Bex, J. Bezshyiko, Ia. Bhom, J. Bieker, M. S. Biesuz, N. V. Billoir, P. Biolchini, A. Birch, M. Bishop, F. C. R. Bitadze, A. Bizzeti, A. Blake, T. Blanc, F. Blank, J. E. Blusk, S. Bocharnikov, V. Boelhauve, J. A. Boente Garcia, O. Boettcher, T. Bohare, A. Boldyrev, A. Bolognani, C. S. Bolzonella, R. Bonacci, R. B. Bondar, N. Bordelius, A. Borgato, F. Borghi, S. Borsato, M. Borsuk, J. T. Bouchiba, S. A. Bovill, M. Bowcock, T. J. V. Boyer, A. Bozzi, C. Brea Rodriguez, A. Breer, N. Brodzicka, J. Brossa Gonzalo, A. Brown, J. Brundu, D. Buchanan, E. Buonaura, A. Buonincontri, L. Burke, A. T. Burr, C. Butter, J. S. Buytaert, J. Byczynski, W. Cadeddu, S. Cai, H. Caillet, A. C. Calabrese, R. Calderon Ramirez, S. Calefice, L. Cali, S. Calvi, M. Calvo Gomez, M. Camargo Magalhaes, P. Cambon Bouzas, J. I. Campana, P. Campora Perez, D. H. Campoverde Quezada, A. F. Capelli, S. Capriotti, L. Nikhef National Institute for Subatomic Physics Amsterdam Netherlands Department of Physics University of Warwick Coventry United Kingdom Physik-Institut Universität Zürich Zürich Switzerland Oliver Lodge Laboratory University of Liverpool Liverpool United Kingdom Syracuse University Syracuse United States Instituto Galego de Física de Altas Enerxías (IGFAE) Universidade de Santiago de Compostela Santiago de Compostela Spain H.H. Wills Physics Laboratory University of Bristol Bristol United Kingdom Department of Physics and Astronomy Uppsala University Uppsala Sweden Université Paris-Saclay CNRS/IN2P3 IJCLab Orsay France University of Michigan Ann Arbor United States Université Clermont Auvergne CNRS/IN2P3 LPC Clermont-Ferrand France University of Cincinnati Cincinnati United States INFN Laboratori Nazionali di Frascati Frascati Italy Fakultät Physik Technische Universität Dortmund Dortmund Germany Lamarr Institute for Machine Learning and Artificial Intelligence Dortmund Germany European Organization for Nuclear Research (CERN) Geneva Switzerland School of Physics and Astronomy University of Glasgow Glasgow United Kingdom Department of Physics and Astronomy University of Manchester Manchester United Kingdom Cavendish Laboratory University of Cambridge Cambridge United Kingdom LPNHE Sorbonne Université Paris Diderot Sorbonne Paris Cité CNRS/IN2P3 Paris France Universidade Federal do Rio de Janeiro (UFRJ) Rio de Janeiro Brazil School of Physics State Key Laboratory of Nuclear Physics and Technology Peking University Beijing China INFN Sezione di Firenze Firenze Italy INFN Sezione di Ferrara Ferrara Italy INFN Sezione di Milano-Bicocca Milano Italy Università degli Studi di Milano-Bicocca Milano Italy University of Chinese Academy of Sciences Beijing China INFN Sezione di Roma Tor Vergata Roma Italy Università di Roma Tor Vergata Roma Italy Consejo Nacional de Rectores (CONARE) San Jose Costa Rica Aix Marseille Univ CNRS/IN2P3 CPPM Mar

The Υ(2S) and Υ(3S) production cross-sections are measured relative to that of the Υ(1S) meson, as a function of charged-particle multiplicity in proton-proton collisions at a centre-of-mass energy of 13 TeV. The measurement uses data collected by the LHCb experiment in 2018 corresponding to an integrated luminosity of 2 fb−1. Both the Υ(2S)-to-Υ(1S) and Υ(3S)-to-Υ(1S) cross-section ratios are found to decrease significantly as a function of event multiplicity, with the Υ(3S)-to-Υ(1S) ratio showing a steeper decline towards high multiplicity. This hierarchy is qualitatively consistent with the comover model predictions, indicating that final-state interactions play an important role in bottomonia production in high-multiplicity events.

关键词：

Visualizing the Energy of Scattered data by Using Cubic Timmer Triangular Patches

学校读者我要写书评

暂无评论

Journal of Physics: Conference Series 2019年第1期1366卷

作者： F A M Ali S A A Karim S C Dass V Skala A Saaban M K Hasan I Hashim Fundamental and Applied Sciences Department Universiti Teknologi PETRONAS 32610 Seri Iskandar Perak Darul Ridzuan Malaysia Fundamental and Applied Sciences Department and Centre for Smart Grid Energy Research (CSMER) Institute of Autonomous System Universiti Teknologi PETRONAS Bandar Seri Iskandar 32610 Seri Iskandar Perak Darul Ridzuan Malaysia School of Computer Science and Engineering University of West Bohemia Plzen School of Quantitative Sciences UUMCAS Universiti Utara Malaysia Kedah Malaysia Centre for Artificial Intelligence Technology Faculty of Information Science and Technology Universiti Kebangsaan Malaysia 43600 UKM Bangi Selangor Malaysia Centre for Modelling & Data Science Faculty of Science & Technology Universiti Kebangsaan Malaysia 43600 UKM Bangi Selangor Malaysia

This paper discusses the application of the new cubic Timmer triangular patches constructed by Ali et al. [1] to interpolate the irregularly scattered data with C 1 continuity. In order to apply the cubic Timmer triangular patches for scattered data interpolation, the data is first triangulated by using the Delaunay algorithm, and then the sufficient condition for C 1 continuity is derived along the adjacent triangles. Two methods will be used to calculate the cubic Timmer ordinates on each triangle. The convex combination between three local schemes Ti , i = 1,2,3 will be used to produce the C 1 surface everywhere. The proposed scheme will be tested to visualize one energy data set with irregular shape properties. Numerical and graphical results are presented by using MATLAB. Comparisons between the proposed scheme and existing schemes such as cubic Ball and cubic Bézier triangular patches are also carried out. The results indicate that the surface produced by cubic Timmer triangular patch is better than surface produced using cubic Ball and cubic Bezier triangular patches with overall coefficient of determination R2 value obtained to be larger than 0.9234.

关键词：

Counting People Based on Linear, Weighted, and Local Random Forests

学校读者我要写书评

暂无评论

Counting People Based on Linear, Weighted, and Local Random ...

Proceedings of the Digital Image Computing: Technqiues and Applications (DICTA)

作者： Helia Farhood Xiangjian He Wenjing Jia Michael Blumenstein Hanhui Li Global Big Data Technologies Centre University of Technology Sydney Sydney Australia Centre for Artificial Intelligence University of Technology Sydney Sydney Australia School of Data and Computer Science Sun Yat-sen University Guangzhou China

Recently, many works have been published for counting people. However, when being applied to real-world train station videos, they have exposed many limitations due to problems such as low resolution, heavy occlusion, various density levels and perspective distortions. In this paper, following the recent trend of regression-based density estimation, we present a linear regression approach based on local Random Forests for counting either standing or moving people on station platforms. By dividing each frame into sub-windows and extracting features with ground truth densities as well as learned weights, we perform a linear transformation for counting people to overcome the perspective problems of the existing patch-based approaches. We present improvements against several recent baselines on the UCSD dataset and a dataset of CCTV videos taken from a train station. We also show improvements in speed compared with the state-of-the-art models based on detection and Deep Learning.

关键词： Feature extraction Estimation Training Videos Linear regression Vegetation Decision trees

A probabilistic atlas of the human thalamic nuclei combining ex vivo MRI and histology

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Iglesias, Juan Eugenio Insausti, Ricardo Lerma-Usabiaga, Garikoitz Bocchetta, Martina van Leemput, Koen Greve, Douglas N. van der Kouwe, Andre Fischl, Bruce Caballero-Gaudes, César Paz-Alonso, Pedro M. Department of Medical Physics and Biomedical Engineering University College London United Kingdom BCBL. Basque Center on Cognition Brain and Language Spain Human Neuroanatomy Laboratory University of Castilla-La Mancha Spain Dementia Research Centre Department of Neurodegenerative Disease Institute of Neurology University College London United Kingdom Martinos Center for Biomedical Imaging Massachusetts General Hospital Harvard Medical School United States Department of Applied Mathematics and Computer Science Technical University of Denmark MIT Computer Science and Artificial Intelligence Laboratory United States

The human thalamus is a brain structure that comprises numerous, highly specific nuclei. Since these nuclei are known to have different functions and to be connected to different areas of the cerebral cortex, it is of great interest for the neuroimaging community to study their volume, shape and connectivity in vivo with MRI. In this study, we present a probabilistic atlas of the thalamic nuclei built using ex vivo brain MRI scans and histological data, as well as the application of the atlas to in vivo MRI segmentation. The atlas was built using manual delineation of 26 thalamic nuclei on the serial histology of 12 whole thalami from six autopsy samples, combined with manual segmentations of the whole thalamus and surrounding structures (caudate, putamen, hippocampus, etc.) made on in vivo brain MR data from 39 subjects. The 3D structure of the histological data and corresponding manual segmentations was recovered using the ex vivo MRI as reference frame, and stacks of blockface photographs acquired during the sectioning as intermediate target. The atlas, which was encoded as an adaptive tetrahedral mesh, shows a good agreement with with previous histological studies of the thalamus in terms of volumes of representative nuclei. When applied to segmentation of in vivo scans using Bayesian inference, the atlas shows excellent test-retest reliability, robustness to changes in input MRI contrast, and ability to detect differential thalamic effects in subjects with Alzheimer’s disease. The probabilistic atlas and companion segmentation tool are publicly available as part of the neuroimaging package FreeSurfer. Copyright © 2018, The Authors. All rights reserved.

关键词： Brain

Towards Automatic Construction of Diverse,High-quality Image datasets

学校读者我要写书评

暂无评论

arXiv 2017年

作者： Yao, Yazhou Zhang, Jian Shen, Fumin Liu, Li Zhu, Fan Zhang, Dongxiang Shen, Heng Tao Inception Institute of Artificial Intelligence Abu Dhabi United Arab Emirates Global Big Data Technologies Center University of Technology Sydney Australia School of Computer Science and Engineering University of Electronic Science and Technology of China

The availability of labeled image datasets has been shown critical for high-level image understanding, which continuously drives the progress of feature designing and models developing. However, constructing labeled image datasets is laborious and monotonous. To eliminate manual annotation, in this work, we propose a novel image dataset construction framework by employing multiple textual queries. We aim at collecting diverse and accurate images for given queries from the Web. Specifically, we formulate noisy textual queries removing and noisy images filtering as a multi-view and multi-instance learning problem separately. Our proposed approach not only improves the accuracy but also enhances the diversity of the selected images. To verify the effectiveness of our proposed approach, we construct an image dataset with 100 categories. The experiments show significant performance gains by using the generated data of our approach on several tasks, such as image classification, cross-dataset generalization, and object detection. The proposed method also consistently outperforms existing weakly supervised and web-supervised approaches. Copyright © 2017, The Authors. All rights reserved.

关键词： Image enhancement

Publisher Correction: The default network of the human brain is associated with perceived social isolation

学校读者我要写书评

暂无评论

Nature communications 2021年第1期12卷 3202页

作者： R Nathan Spreng Emile Dimas Laetitia Mwilambwe-Tshilobo Alain Dagher Philipp Koellinger Gideon Nave Anthony Ong Julius M Kernbach Thomas V Wiecki Tian Ge Yue Li Avram J Holmes B T Thomas Yeo Gary R Turner Robin I M Dunbar Danilo Bzdok Laboratory of Brain and Cognition Montreal Neurological Institute Department of Neurology and Neurosurgery Faculty of Medicine McGill University Montreal QC Canada. nathan.spreng@***. Departments of Psychiatry and Psychology McGill University Montreal QC Canada. nathan.spreng@***. Douglas Mental Health University Institute Verdun QC HRH 1R3 Canada. nathan.spreng@***. McConnell Brain Imaging Centre Montreal Neurological Institute (MNI) McGill University Montreal QC Canada. nathan.spreng@***. Department of Biomedical Engineering Faculty of Medicine McGill University Montreal QC Canada. Laboratory of Brain and Cognition Montreal Neurological Institute Department of Neurology and Neurosurgery Faculty of Medicine McGill University Montreal QC Canada. McConnell Brain Imaging Centre Montreal Neurological Institute (MNI) McGill University Montreal QC Canada. School of Business and Economics Vrije Universiteit Amsterdam Amsterdam The Netherlands. Marketing Department the Wharton School University of Pennsylvania Pennsylvania PA USA. Department of Human Development Cornell University Ithaca NY USA. Division of Geriatrics and Palliative Medicine Weill Cornell Medical College New York NY USA. Department of Neurosurgery Neurosurgical Artificial Intelligence Laboratory Aachen (NAILA) RWTH Aachen University Hospital Aachen Germany. Quantopian Inc. Boston MA USA. Psychiatric and Neurodevelopmental Genetics Unit Center for Genomic Medicine Massachusetts General Hospital Boston MA 02114 USA. Department of Psychiatry Massachusetts General Hospital Harvard Medical School Boston MA 02114 USA. Athinoula A. Martinos Center for Biomedical Imaging Massachusetts General Hospital Charlestown MA 02129 USA. Stanley Center for Psychiatric Research Broad Institute of MIT and Harvard Cambridge MA 02138 USA. School of Computer Science McGill University Montreal QC Canada. Departments of Psychology and Psychiatry Yale University New Haven CA 06520 USA. Depa

Learning student networks via feature embedding

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Chen, Hanting Wang, Yunhe Xu, Chang Xu, Chao Tao, Dacheng Coopertative Medianet Innovation Center School of EECS Peking University Beijing100871 China Noah’s Ark Laboratory Huawei Technologies Co. Ltd HuaWei Building No.3 Xinxi Road Shang-Di Information Industri Base Hai-Dian District Beijing100085 China UBTech Sydney Artificial Intelligence Centre School of Computer Science Faculty of Engineering and Information Technologies University of Sydney J12 Cleveland St DarlingtonNSW2008 Australia

—Deep convolutional neural networks have been widely used in numerous applications, but their demanding storage and computational resource requirements prevent their applications on mobile devices. Knowledge distillation aims to optimize a portable student network by taking the knowledge from a well-trained heavy teacher network. Traditional teacher-student based methods used to rely on additional fully-connected layers to bridge intermediate layers of teacher and student networks, which brings in a large number of auxiliary parameters. In contrast, this paper aims to propagate information from teacher to student without introducing new variables which need to be optimized. We regard the teacher-student paradigm from a new perspective of feature embedding. By introducing the locality preserving loss, the student network is encouraged to generate the low-dimensional features which could inherit intrinsic properties of their corresponding high-dimensional features from teacher network. The resulting portable network thus can naturally maintain the performance as that of the teacher network. Theoretical analysis is provided to justify the lower computation complexity of the proposed method. Experiments on benchmark datasets and well-trained networks suggest that the proposed algorithm is superior to state-of-the-art teacher-student learning methods in terms of computational and storage complexity. Copyright © 2018, The Authors. All rights reserved.

关键词： Deep neural networks