检索结果-内蒙古大学图书馆

arXiv 2018年

作者： Cui, Guoxin Xu, Jun Zeng, Wei Lan, Yanyan Guo, Jiafeng Cheng, Xueqi CAS Key Laboratory of Network Data Science and Technology Institute of ComputingTechnology Chinese Academy of Sciences Beijing100190 China

One of the most significant bottleneck in training large scale machine learning models on parameter server (PS) is the communication overhead, because it needs to frequently exchange the model gradients between the workers and servers during the training iterations. Gradient quantization has been proposed as an effective approach to reducing the communication volume. One key issue in gradient quantization is setting the number of bits for quantizing the gradients. Small number of bits can significantly reduce the communication overhead while hurts the gradient accuracies, and vise versa. An ideal quantization method would dynamically balance the communication overhead and model accuracy, through adjusting the number bits according to the knowledge learned from the immediate past training iterations. Existing methods, however, quantize the gradients either with fixed number of bits, or with predefined heuristic rules. In this paper we propose a novel adaptive quantization method within the framework of reinforcement learning. The method, referred to as MQGrad, formalizes the selection of quantization bits as actions in a Markov decision process (MDP) where the MDP states records the information collected from the past optimization iterations (e.g., the sequence of the loss function values). During the training iterations of a machine learning algorithm, MQGrad continuously updates the MDP state according to the changes of the loss function. Based on the information, MDP learns to select the optimal actions (number of bits) to quantize the gradients. Experimental results based on a benchmark dataset showed that MQGrad can accelerate the learning of a large scale deep neural network while keeping its prediction accuracies. Copyright © 2018, The Authors. All rights reserved.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

Relative Importance Sampling for off-Policy Actor-Critic in Deep Reinforcement Learning

arXiv

引用

arXiv 2018年

作者： Humayoo, Mahammad Cheng, Xueqi CAS Key Laboratory of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences Beijing100190 China University of Chinese Academy of Sciences Beijing100190 China

Off-policy learning is more unstable compared to on-policy learning in reinforcement learning (RL). One of the reasons for instability of off-policy learning is a discrepancy between target (π) and behavior (b) policy distribution. The discrepancy between π and b distribution can be alleviated by employing the smooth variant of importance sampling (IS), such as relative importance sampling (RIS).The RIS has parameter β ∈ [0, 1] that controls the smoothness. To cope with the instability of off-policy learning, we present the first relative importance sampling-off-policy actor-critic (RIS-off-PAC) model-free algorithms in RL. In our method, A network yields the target policy (actor), the value function (critic) assessing the current policy (π) using samples drawn from the behavior policy. We use action values generated from the behavior policy in reward function to train our algorithms rather than from the target policy. We also use deep neural networks to train both the actor and critic. We evaluated our algorithms on a number of OpenAI Gym problems and demonstrated better or comparable performance to several state-of-the-art RL baselines. Copyright © 2018, The Authors. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

General-Purpose Quantum Circuit Simulator with Projected Entangled-Pair States and the Quantum Supremacy Frontier

引用

Physical Review Letters 2019年第19期123卷 190501-190501页

作者： Chu Guo Yong Liu Min Xiong Shichuan Xue Xiang Fu Anqi Huang Xiaogang Qiang Ping Xu Junhua Liu Shenggen Zheng He-Liang Huang Mingtang Deng Dario Poletti Wan-Su Bao Junjie Wu Henan Key Laboratory of Quantum Information and Cryptography IEU Zhengzhou 450001 China Institute for Quantum Information & State Key Laboratory of High Performance Computing College of Computer National University of Defense Technology Changsha 410073 China Information Systems Technology and Design Singapore University of Technology and Design 8 Somapah Road 487372 Singapore Quantum Intelligence Lab (QI-Lab) Supremacy Future Technologies (SFT) Guangzhou 511340 China Center for Quantum Computing Peng Cheng Laboratory Shenzhen 518055 China Hefei National Laboratory for Physical Sciences at Microscale and Department of Modern Physics University of Science and Technology of China Hefei Anhui 230026 China CAS Centre for Excellence and Synergetic Innovation Centre in Quantum Information and Quantum Physics University of Science and Technology of China Hefei Anhui 230026 China Science and Math Cluster and EPD Pillar Singapore University of Technology and Design 8 Somapah Road 487372 Singapore

Recent advances on quantum computing hardware have pushed quantum computing to the verge of quantum supremacy. Here, we bring together many-body quantum physics and quantum computing by using a method for strongly interacting two-dimensional systems, the projected entangled-pair states, to realize an effective general-purpose simulator of quantum algorithms. The classical computing complexity of this simulator is directly related to the entanglement generation of the underlying quantum circuit rather than the number of qubits or gate operations. We apply our method to study random quantum circuits, which allows us to quantify precisely the memory usage and the time requirements of random quantum circuits. We demonstrate our method by computing one amplitude for a 7×7 lattice of qubits with depth (1+40+1) on the Tianhe-2 supercomputer.

关键词： Quantum benchmarking Quantum computation Quantum simulation Computational complexity Tensor network methods

来源：评论

学校读者我要写书评

暂无评论

Investigating EEG-Based Functional Connectivity Patterns for Multimodal Emotion Recognition

arXiv

引用

arXiv 2020年

作者： Wu, Xun Zheng, Wei-Long Lu, Bao-Liang Center for Brain-Like Computing and Machine Intelligence Department of Computer Science and Engineering Key Laboratory of Shanghai Education Commission for Intelligent Interaction and Cognitive Engineering Brain Science and Technology Research Center Qing Yuan Research Institute Shanghai Jiao Tong University 800 Dong Chuan Road Shanghai200240 China Clinical Data Animation Center Department of Neurology Massachusetts General Hospital Harvard Medical School 55 Fruit Street BostonMA United States

Compared with the rich studies on the motor brain-computer interface (BCI), the recently emerging affective BCI presents distinct challenges since the brain functional connectivity networks involving emotion are not well investigated. Previous studies on emotion recognition based on electroencephalography (EEG) signals mainly rely on single-channel-based feature extraction methods. In this paper, we propose a novel emotion-relevant critical subnetwork selection algorithm and investigate three EEG functional connectivity network features: strength, clustering coefficient, and eigenvector centrality. The discrimination ability of the EEG connectivity features in emotion recognition is evaluated on three public emotion EEG datasets: SEED, SEED-V, and DEAP. The strength feature achieves the best classification performance and outperforms the state-of-the-art differential entropy feature based on single-channel analysis for the EEG signals. The experimental results reveal that distinct functional connectivity patterns are exhibited for the five emotions of disgust, fear, sadness, happiness, and neutrality. Furthermore, we construct a multimodal emotion recognition model by combining the functional connectivity features from EEG and the features from eye movements or physiological signals using deep canonical correlation analysis. The classification accuracies of multimodal emotion recognition are 95.08 ± 6.42% on the SEED dataset, 84.51 ± 5.11% on the SEED-V dataset, and 85.34 ± 2.90% and 86.61 ± 3.76% for arousal and valence on the DEAP dataset, respectively. The results demonstrate the complementary representation properties of the EEG functional connectivity network features with eye movement data. In addition, we find that the brain networks constructed with fewer channels, i.e., 18 channels, achieve comparable performance with that of the 62-channel network with respect to multimodal emotion recognition and enable easier setups for BCI systems in real scenarios. Cop

关键词： Electrophysiology

来源：评论

学校读者我要写书评

暂无评论

Robust classification with sparse representation fusion on diverse data subsets

arXiv

引用

arXiv 2019年

作者： Feng, Chun-Mei Xu, Yong Li, Zuoyong Yang, Jian Bio-Computing Research Center Harbin Institute of Technology Shenzhen Guangdong518055 China Key Laboratory of Network Oriented Intelligent Computation Shenzhen Guangdong518055 China Fuzhou350121 China School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing Jiangsu210094 China

Sparse Representation (SR) techniques encode the test samples into a sparse linear combination of all training samples and then classify the test samples into the class with the minimum residual. The classification of SR techniques depends on the representation capability on the test samples. However, most of these models view the representation problem of the test samples as a deterministic problem, ignoring the uncertainty of the representation. The uncertainty is caused by two factors, random noise in the samples and the intrinsic randomness of the sample set, which means that if we capture a group of samples, the obtained set of samples will be different in different conditions. In this paper, we propose a novel method based upon Collaborative Representation that is a special instance of SR and has closed-form solution. It performs Sparse Representation Fusion based on the Diverse Subset of training samples (SRFDS), which reduces the impact of randomness of the sample set and enhances the robustness of classification results. The proposed method is suitable for multiple types of data and has no requirement on the pattern type of the tasks. In addition, SRFDS not only preserves a closed-form solution but also greatly improves the classification performance. Promising results on various datasets serve as the evidence of better performance of SRFDS than other SR-based methods. The Matlab code of SRFDS will be accessible at http://***/***. Copyright © 2019, The Authors. All rights reserved.

关键词： Sampling

来源：评论

学校读者我要写书评

暂无评论

Constraining Cosmological Phase Transitions with the Parkes Pulsar Timing Array

引用

Physical Review Letters 2021年第25期127卷 251303-251303页

作者： Xiao Xue Ligong Bian Jing Shu Qiang Yuan Xingjiang Zhu N. D. Ramesh Bhat Shi Dai Yi Feng Boris Goncharov George Hobbs Eric Howard Richard N. Manchester Christopher J. Russell Daniel J. Reardon Ryan M. Shannon Renée Spiewak Nithyanandan Thyagarajan Jingbo Wang CAS Key Laboratory of Theoretical Physics Institute of Theoretical Physics Chinese Academy of Sciences Beijing 100190 China School of Physical Sciences University of Chinese Academy of Sciences Beijing 100049 China II. Institute of Theoretical Physics Universität Hamburg 22761 Hamburg Germany Department of Physics Chongqing University Chongqing 401331 China Chongqing Key Laboratory for Strongly Coupled Physics Chongqing 401331 China School of Fundamental Physics and Mathematical Sciences Hangzhou Institute for Advanced Study University of Chinese Academy of Sciences Hangzhou 310024 China Center for High Energy Physics Peking University Beijing 100871 China International Center for Theoretical Physics Asia-Pacific Beijing/Hanzhou China Key Laboratory of Dark Matter and Space Astronomy Purple Mountain Observatory Chinese Academy of Sciences Nanjing 210023 China School of Astronomy and Space Science University of Science and Technology of China Hefei 230026 China School of Physics and Astronomy Monash University Clayton VIC 3800 Australia OzGrav: The ARC Centre of Excellence for Gravitational Wave Discovery Hawthorn VIC 3122 Australia Advanced Institute of Natural Sciences Beijing Normal University at Zhuhai 519087 China International Centre for Radio Astronomy Research Curtin University Bentley WA 6102 Australia Western Sydney University Locked Bag 1797 Penrith South DC NSW 1797 Australia University of Chinese Academy of Sciences Beijing 100049 China CSIRO Astronomy and Space Science P.O. Box 76 Epping NSW 1710 Australia Macquarie University Department of Physics and Astronomy Sydney NSW 2109 Australia CSIRO Scientific Computing Australian Technology Park Locked Bag 9013 Alexandria NSW 1435 Australia Centre for Astrophysics and Supercomputing Swinburne University of Technology P.O. Box 218 Hawthorn VIC 3122 Australia Jodrell Bank Centre for Astrophysics University of Manchester Manchester M13 9PL United Kingdom CSIRO Astronomy and Space Science

A cosmological first-order phase transition is expected to produce a stochastic gravitational wave background. If the phase transition temperature is on the MeV scale, the power spectrum of the induced stochastic gravitational waves peaks around nanohertz frequencies, and can thus be probed with high-precision pulsar timing observations. We search for such a stochastic gravitational wave background with the latest data set of the Parkes Pulsar Timing Array. We find no evidence for a Hellings-Downs spatial correlation as expected for a stochastic gravitational wave background. Therefore, we present constraints on first-order phase transition model parameters. Our analysis shows that pulsar timing is particularly sensitive to the low-temperature (T∼1–100 MeV) phase transition with a duration (β/H*)−1∼10−2−10−1 and therefore can be used to constrain the dark and QCD phase transitions.

关键词： First order phase transitions Gravitational wave detection

来源：评论

学校读者我要写书评

暂无评论

Dynamic-K recommendation with personalized decision boundary 23rd

Dynamic-K recommendation with personalized decision boundary

引用

23rd China conference on Information Retrieval, CCIR 2017

作者： Gao, Yan Guo, Jiafeng Lan, Yanyan Liao, Huaming CAS Key Lab of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences Beijing China

ISBN: (纸本)9783319686981

In this paper, we investigate the recommendation task in the most common scenario with implicit feedback (e.g., clicks, purchases). State-of-the-art methods in this direction usually cast the problem as to learn a personalized ranking on a set of items (e.g., webpages, products). The top-N results are then provided to users as recommendations, where the N is usually a fixed number pre-defined by the system according to some heuristic criteria (e.g., page size, screen size). There is one major assumption underlying this fixed-number recommendation scheme, i.e., there are always sufficient relevant items to users’ preferences. Unfortunately, this assumption may not always hold in real-world scenarios. In some applications, there might be very limited candidate items to recommend, and some users may have very high relevance requirement in recommendation. In this way, even the top-1 ranked item may not be relevant to a user’s preference. Therefore, we argue that it is critical to provide a dynamic-K recommendation, where the K should be different with respect to the candidate item set and the target user. We formulate this dynamic-K recommendation task as a joint learning problem with both ranking and classification objectives. The ranking objective is the same as existing methods, i.e., to create a ranking list of items according to users’ interests. The classification objective is unique in this work, which aims to learn a personalized decision boundary to differentiate the relevant items from irrelevant items. Based on these ideas, we extend two state-of-the-art ranking-based recommendation methods, i.e., BPRMF and HRM, to the corresponding dynamic-K versions, namely DK-BPRMF and DK-HRM. Our experimental results on two datasets show that the dynamic-K models are more effective than the original fixed-N recommendation methods. © Springer International Publishing AG 2017.

关键词： Websites

来源：评论

学校读者我要写书评

暂无评论

Beyond precision: A study on recall of initial retrieval with neural representations

arXiv

引用

arXiv 2018年

作者： Xiao, Yan Guo, Jiafeng Fan, Yixing Lan, Yanyan Xu, Jun Cheng, Xueqi University of Chinese Academy of Sciences Beijing China CAS Key Lab of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences Beijing China

Vocabulary mismatch is a central problem in information retrieval (IR), i.e., the relevant documents may not contain the same (symbolic) terms of the query. Recently, neural representations have shown great success in capturing semantic relatedness, leading to new possibilities to alleviate the vocabulary mismatch problem in IR. However, most existing efforts in this direction have been devoted to the re-ranking stage. That is to leverage neural representations to help re-rank a set of candidate documents, which are typically obtained from an initial retrieval stage based on some symbolic index and search scheme (e.g., BM25 over the inverted index). This naturally raises a question: If the relevant documents have not been found in the initial retrieval stage due to vocabulary mismatch, there would be no chance to re-rank them to the top positions later. Therefore, in this paper, we study the problem how to employ neural representations to improve the recall of relevant documents in the initial retrieval stage. Specifically, to meet the efficiency requirement of the initial stage, we introduce a neural index for the neural representations of documents, and propose two hybrid search schemes based on both neural and symbolic indices, namely the parallel search scheme and the sequential search scheme. Our experiments show that both hybrid index and search schemes can improve the recall of the initial retrieval stage with small overhead. Copyright © 2018, The Authors. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Fully-convolutional intensive feature flow neural network for text recognition

arXiv

引用

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

来源：评论

学校读者我要写书评

暂无评论

Coarse to fine: Diffusing categories in wikipedia 26

Coarse to fine: Diffusing categories in wikipedia

引用

26th International World Wide Web Conference, WWW 2017 Companion

作者： Cai, Pengshan Feng, Yansong Jia, Yantao Wang, Yuanzhuo Jin, Xiaolong Cheng, Xueqi CAS Key Laboratory of Network Data Science and Technology Institute of Computing Technology Chinese Academy of Sciences Beijing China Institute of Computer Science and Technology Peking University Beijing China

ISBN: (纸本)9781450349147

Automatic taxonomy construction aims to build a categorization system without human efforts. Traditional textual pattern based methods extract hyponymy relation in raw texts. However, these methods usually yield low precision and recall. In this paper, we propose a method to automatically find diffusing attributes to a category from Wikipedia infoboxes. We use the diffusing attribute to diffuse a coarse-grained category into several fine-grained subcategories and generate a finer-grained taxonomy. Experiments show our method can find proper diffusing attributes to categories across various domains. © 2017 International World Wide Web Conference Committee (IW3C2), published under Creative Commons CC BY 4.0 License.

关键词： Taxonomies

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：