检索结果-内蒙古大学图书馆

A decentralized Nesterov gradient method for stochastic optimization over unbalanced directed networks

ASIAN JOURNAL OF CONTROL 2022年第2期24卷 576-593页

作者： Hu, Jinhui Xia, Dawen Cheng, Huqiang Feng, Liping Ji, Lianghao Guo, Jing Li, Huaqing Southwest Univ Chongqing Key Lab Nonlinear Circuits & Intelligen Coll Elect & Informat Engn 2 Tiansheng Rd Chongqing Peoples R China Guizhou Minzu Univ Coll Data Sci & Informat Engn Guiyang Peoples R China Xinzhou Teachers Univ Dept Comp Sci Xinzhou Shanxi Peoples R China Chongqing Univ Posts & Telecommun Chongqing Key Lab Computat Intelligence Chongqing Peoples R China

Decentralized stochastic gradient methods play significant roles in large-scale optimization that finds many practical applications in machine learning and coordinated control. This paper studies optimization problems over unbalanced directed networks, where the mutual goal of agents in the network is to optimize a global objective function expressed as a sum of local objective functions. Each agent using only local computation and communication in the networks is assumed to get access to a stochastic first-order oracle. In order to devise a noise-tolerant decentralized algorithm with accelerated linear convergence, a decentralized Nesterov gradient algorithm with the constant step-size and parameter using stochastic gradients is proposed in this paper. The proposed algorithm employing a gradient-tracking technique is proved to converge linearly to an error ball around the optimal solution via the analysis on a linear system when the positive constant step-size and parameter are sufficiently small. We further recover the exact linear convergence for the proposed algorithm with exact gradients under the same selection conditions of the constant step-size and parameter. Some real-world data sets are used in simulations to validate the correctness of the theoretical findings and practicability of the proposed algorithm.

关键词： decentralized optimization unbalanced directed networks stochastic gradients machine learning multi‐ agent systems

来源：评论

学校读者我要写书评

暂无评论

Video Based Cocktail Causal Container for Blood Pressure Classification and Blood Glucose Prediction

引用

IEEE JOURNAL OF BIOMEDICAL AND HEALTH informatICS 2023年第2期27卷 1118-1128页

作者： Zhang, Chuanhao Jovanov, Emil Liao, Hongen Zhang, Yuan-Ting Lo, Benny Zhang, Yuan Guan, Cuntai Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligent Chongqing 400715 Peoples R China Univ Alabama Dept Elect & Comp Engn Huntsville AL 35899 USA Tsinghua Univ Sch Med Dept Biomed Engn Beijing 100190 Peoples R China Hong Kong Ctr Cerebro Cardiovasc Hlth Engn Hong Kong Peoples R China City Univ Hong Kong Dept Biomed Engn Hong Kong Peoples R China Imperial Coll London Dept Surg & Canc London SW7 2BX England Nanyang Technol Univ S Lab Singapore 639798 Singapore

With the development of modern cameras, more physiological signals can be obtained from portable devices like smartphone. Some hemodynamically based non-invasive video processing applications have been applied for blood pressure classification and blood glucose prediction objectives for unobtrusive physiological monitoring at home. However, this approach is still under development with very few publications. In this paper, we propose an end-to-end framework, entitled cocktail causal container, to fuse multiple physiological representations and to reconstruct the correlation between frequency and temporal information during multi-task learning. Cocktail causal container processes hematologic reflex information to classify blood pressure and blood glucose. Since the learning of discriminative features from video physiological representations is quite challenging, we propose a token feature fusion block to fuse the multi-view fine-grained representations to a union discrete frequency space. A causal net is used to analyze the fused higher-order information, so that the framework can be enforced to disentangle the latent factors into the related endogenous association that corresponds to down-stream fusion information to improve the semantic interpretation. Moreover, a pair-wise temporal frequency map is developed to provide valuable insights into extraction of salient photoplethysmograph (PPG) information from fingertip videos obtained by a standard smart phone camera. Extensive comparisons have been implemented for the validation of cocktail causal container using a Clinical dataset and PPG-BP benchmark. The root mean square error of 1.329 +/- 0.167 for blood glucose prediction and precision of 0.89 +/- 0.03 for blood pressure classification are achieved in Clinical dataset.

关键词： Feature extraction Blood Task analysis Multitasking Biomedical monitoring Glucose Deep learning Blood glucose prediction blood pressure classification deep learning non-invasive measurement

来源：评论

学校读者我要写书评

暂无评论

Integration of high-κ native oxides of gallium for two-dimensional transistors

引用

NATURE ELECTRONICS 2024年第12期7卷 1126-1136页

作者： Yi, Kongyang Qin, Wen Huang, Yamin Wu, Yao Feng, Shaopeng Fang, Qiyi Cao, Xun Deng, Ya Zhu, Chao Zou, Xilu Ang, Kah-Wee Li, Taotao Wang, Xinran Lou, Jun Lai, Keji Hu, Zhili Zhang, Zhuhua Dong, Yemin Kalantar-Zadeh, Kourosh Liu, Zheng Nanyang Technol Univ Sch Mat Sci & Engn Singapore Singapore Nanjing Univ Aeronaut & Astronaut State Key Lab Mech & Control Aerosp Struct Key Lab Intelligent Nano Mat & Devices Minist Educ Nanjing Peoples R China Nanjing Univ Aeronaut & Astronaut Inst Frontier Sci Nanjing Peoples R China Chinese Acad Sci Shanghai Inst Microsyst & Informat Technol State Key Lab Mat Integrated Circuits Shanghai Peoples R China Univ Texas Austin Dept Phys Austin TX 78712 USA Rice Univ Dept Mat Sci & NanoEngn Houston TX USA Nanjing Univ Sch Elect Sci & Engn Nanjing Peoples R China Suzhou Univ Suzhou Peoples R China Natl Univ Singapore Dept Elect & Comp Engn Singapore Singapore Nanjing Univ Sch Integrated Circuits Suzhou Peoples R China Nanjing Univ Interdisciplinary Res Ctr Future Intelligent Chips Suzhou Peoples R China Univ Sydney Sch Chem & Biomol Engn Sydney NSW Australia CINTRA CNRS NTU THALES Singapore Singapore

The deposition of a metal oxide layer with good dielectric properties is a critical step in fabricating the gate dielectric of transistors based on two-dimensional semiconductors. However, current techniques for depositing ultrathin metal oxide layers on two-dimensional semiconductors suffer from quality issues that can compromise transistor performance. Here, we show that an ultrathin and uniform native oxide of gallium (Ga2O3) that naturally forms on the surface of liquid metals in an ambient environment can be prepared on the surface of molybdenum disulfide (MoS2) by squeeze-printing and surface-tension-driven methods. The Ga2O3 layer possesses a high dielectric constant of around 30 and equivalent oxide thickness of around 0.4 nm. Due to the good dielectric properties and van der Waals integration, MoS2 transistors with Ga2O3 gate dielectrics exhibit a subthreshold swing down to 60 mV dec-1, an on/off ratio of 108 and a gate leakage down to around 4 x 10-7 A cm-2. Ultrathin films of gallium oxide with a dielectric constant of around 30 can be formed on the surface of molybdenum disulfide using a liquid metal-based approach and used as the gate insulator in transistors.

关键词： Van der Waals forces

来源：评论

学校读者我要写书评

暂无评论

A Network Intrusion Detection System with Broadband WO3-x/WO3-x-Ag/WO3-x Optoelectronic Memristor

引用

ADVANCED FUNCTIONAL MATERIALS 2024年第23期34卷 2312885-2312885页

作者： Yang, Wenhao Kan, Hao Shen, Guozhen Li, Yang Univ Jinan Sch Informat Sci & Engn Shandong Prov Key Lab Network Based Intelligent Co Jinan 250022 Peoples R China Beijing Inst Technol Sch Integrated Circuits & Elect Beijing 100081 Peoples R China Shandong Univ Sch Integrated Circuits Jinan 250101 Peoples R China

Real-time intrusion detection system based on the von Neumann architecture struggle to balance low power consumption and high computing speed. In this work, a strategy for network intrusion detection system based on the WO3-x/WO3-x-Ag/WO3-x structured optoelectronic memristor overcoming the aforementioned issues is proposed and demonstrated. Through the modulation of electrical signals, the memristor successfully simulates a series of important synaptic functionalities including short-term/long-term synaptic plasticity. Meanwhile, when subjected to light stimulus, it demonstrates remarkable synaptic behaviors in terms of long/short-term memory and "learning-forgetting-relearning." Based on this memristor array, a convolutional neural network is constructed to recognize abnormal network records within the KDDCup-99 dataset accurately and efficiently. The power consumption (10(-6) W) is over seven orders of magnitude lower than that of central processing unit, etc. Subsequently, an intrusion detection system is established to integrate collection, processing, and detection of real-time network data, successfully classifying various types of network records. Hence, this work is expected to promote the development of high-density storage and neuromorphic computing technology, and provides an application idea for intelligent electronic devices.

关键词： intrusion detection system KDDCup-99 dataset neural network optoelectronic memristor WO3-x

来源：评论

学校读者我要写书评

暂无评论

Analysis of autonomic nervous patterns before and during sleep deprivation 2021

Analysis of autonomic nervous patterns before and during sle...

引用

2nd International Conference on Artificial Intelligence and information Systems (ICAIIS )

作者： Yang, Xuehong Kong, Fanmeng Xiong, Ronglong Liu, Guangyuan Wen, Wanhui Southwest Univ Sch Elect & Informat Engn Chongqing 400715 Peoples R China Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China

ISBN: (纸本)9781450390200

Sleep deprivation (SD) is harmful to human health. This paper analyzed the autonomic nervous patterns (ANP) of subjects before sleep deprivation (BSD) and during sleep deprivation (DSD). Electrocardiogram (ECG) data were obtained from 60 subjects, and data were collected for half an hour before and during SD for all subjects. Twenty-five heart rate variability features were collected from RR intervals of ECG data for classification. The results of classification showed that BSD and DSD were distinguishable in their autonomic nervous system (ANS) activities. The optimal feature subset was selected by the sequential backward selection (SBS) algorithm, and the validation accuracy of BSD and DSD classification of the support vector machine (SVM) was 80% by using the combination of linear features and nonlinear features of the RR interval series.

关键词： Sleep deprivation autonomic nervous system heart rate variability classification

来源：评论

学校读者我要写书评

暂无评论

An Efficient nonlinear Dichotomous Coordinate Descent Adaptive Algorithm Based on Random Fourier Features

引用

IEEE SIGNAL PROCESSING LETTERS 2020年 27卷 1804-1808页

作者： Chen, Tao Wang, Shiyuan Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China

The auxiliary normal equation is proposed to construct an incremental update (IU) system in which the increment of weight vector rather than the weight itself is optimized at each iteration, which however, can only deal with problems under linearity assumption. In this letter, we establish the IU system in a fixed-dimension feature space induced by the random Fourier features mapping, enabling it to address nonlinear issues efficiently. Then, to obtain the solution vector to the IU system and overcome high complexity and instability incurred by matrix inversion, a dichotomous coordinate descent (DCD) method is employed to propose an efficient random Fourier features kernel dichotomous coordinate descent (RFFKDCD) algorithm. The proposed RFFKDCD algorithm can tackle nonlinear issues efficiently while avoiding high computational complexity existing widely in kernel-based algorithms. Simulations on both synthetic and real-world examples show that the proposed RFFKDCD algorithm outperforms the recursive least squares-based algorithms in terms of computational load.

关键词： Signal processing algorithms Kernel Mathematical model Computational complexity Convergence Optimization Time series analysis Normal equation incremental update random Fourier features dichotomous coordinate descent

来源：评论

学校读者我要写书评

暂无评论

A New Settling-time Estimation Protocol to Finite-time Synchronization of Impulsive Memristor-Based Neural Networks

引用

IEEE TRANSACTIONS ON CYBERNETICS 2022年第6期52卷 4312-4322页

作者： Wang, Xin Park, Ju H. Yang, Huilan Zhong, Shouming Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuits & Intelligen Chongqing 400715 Peoples R China Yeungnam Univ Dept Elect Engn Gyongsan 38541 South Korea Southwest Univ Sch Math & Stat Chongqing 400715 Peoples R China Univ Elect Sci & Technol China Sch Math Sci Chengdu 611731 Peoples R China

In this article, the issues of finite-time synchronization and finite-time adaptive synchronization for the impulsive memristive neural networks (IMNNs) with discontinuous activation functions (DAFs) and hybrid impulsive effects are probed into and elaborated on, where the stabilizing impulses (SIs), inactive impulses (IIs), and destabilizing impulses (DIs) are taken into account, respectively. Not resembling several earlier works, a more extensive range of impulses in the context of impulsive effects has been analyzed without using the known average impulsive interval strategy (AIIS). In light of the theories of differential inclusions and set-valued map, as well as impulsive control, new sufficient criteria with respect to the estimated settling time for synchronization of the related IMNNs are established using two types of switching control approaches, which sufficiently utilize information from not only the SIs, DIs, and DAFs but also the impulse sequences. Two simulation experiments are presented to the efficiency of the proposed results.

关键词： Synchronization Silicon Switches Protocols Multi-layer neural network Adaptive synchronization destabilizing impulses (DIs) memristive neural networks (MNNs) settling time

来源：评论

学校读者我要写书评

暂无评论

Delay-Dependent Stability Analysis for Switched Stochastic Networks With Proportional Delay

引用

IEEE TRANSACTIONS ON CYBERNETICS 2022年第7期52卷 6369-6378页

In this article, the issue of exponential stability (ES) is investigated for a class of switched stochastic neural networks (SSNNs) with proportional delay (PD). The key feature of PD is an unbounded time-varying delay. By considering the comparison principle and combining the extended formula for the variation of parameters, we conquer the difficulty in consideration of PD effects for such networks for the first time, where the subsystems addressed may be stable or unstable. New delay-dependent conditions with respect to the mean-square ES of systems are established by employing the average dwell-time (ADT) technique, stochastic analysis theory, and Lyapunov approach. It is shown that the acquired minimum average dwell time (MADT) is not only relevant to the stable subsystems (SSs) and unstable subsystems (USs) but also dependent on the decay ratio (DR), increasing ratio (IR), as well as PD. Finally, the availability of the derived results under an average dwell-time-switched regulation (ADTSR) is illustrated through two numerical simulation examples.

关键词： Switches Artificial neural networks Delays Stability criteria Asymptotic stability Synchronization Switched systems Comparison principle proportional delay (PD) switched neural networks (NNs) unstable subsystems (USs)

来源：评论

学校读者我要写书评

暂无评论

An index-free sparse neural network using two-dimensional semiconductor ferroelectric field-effect transistors

引用

NATURE ELECTRONICS 2025年第3期8卷 1-13页

作者： Ning, Hongkai Wen, Hengdi Meng, Yuan Yu, Zhihao Fu, Yuxiang Zou, Xilu Shen, Yilin Luo, Xiai Zhao, Qiyue Zhang, Tao Liu, Lei Zhu, Shitong Li, Taotao Li, Weisheng Li, Li Gao, Li Shi, Yi Wang, Xinran Nanjing Univ Sch Elect Sci & Engn Natl Lab Solid State Microstruct Nanjing Peoples R China Nanjing Univ Collaborat Innovat Ctr Adv Microstruct Nanjing Peoples R China Nanjing Univ Posts & Telecommun Sch Integrated Circuit Sci & Engn Nanjing Peoples R China Suzhou Lab Cent Lab Suzhou Peoples R China Nanjing Univ Sch Integrated Circuits Suzhou Peoples R China Nanjing Univ Interdisciplinary Res Ctr Future Intelligent Chips Suzhou Peoples R China Nanjing Univ Posts & Telecommun Inst Adv Mat Sch Mat Sci & Engn State Key Lab Organ Elect & Informat Displays Nanjing Peoples R China

The fine-grained dynamic sparsity in biological synapses is an important element in the energy efficiency of the human brain. Emulating such sparsity in an artificial system requires off-chip memory indexing, which has a considerable energy and latency overhead. Here, we report an in-memory sparsity architecture in which index memory is moved next to individual synapses, creating a sparse neural network without external memory indexing. We use a compact building block consisting of two non-volatile ferroelectric field-effect transistors acting as a digital sparsity and an analogue weight. The network is formulated as the Hadamard product of the sparsity and weight matrices, and the hardware, which is comprised of 900 ferroelectric field-effect transistors, is based on wafer-scale chemical-vapour-deposited molybdenum disulfide integrated through back-end-of-line processes. With the system, we demonstrate key synaptic processes-including pruning, weight update and regrowth-in an unstructured and fine-grained manner. We also develop a vectorial approximate update algorithm and optimize training scheduling. Through this software-hardware co-optimization, we achieve 98.4% accuracy in an EMNIST letter recognition task under 75% sparsity. Simulations on large neural networks show a tenfold reduction in latency and a ninefold reduction in energy consumption when compared with a dense network of the same performance.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

Investigation on the combined model of sensor drift compensation and open-set gas recognition based on electronic nose datasets

引用

CHEMOMETRICS AND intelligent labORATORY SYSTEMS 2023年 242卷

作者： Yao, Youbin Chen, Bin Liu, Chuanjun Qu, Cheng Southwest Univ Coll Elect & Informat Engn Chongqing Key Lab Nonlinear Circuit & Intelligent Chongqing 400715 Peoples R China Kyushu Univ Grad Sch Informat Sci & Elect Engn Dept Elect Fukuoka 8190395 Japan

Electronic nose (EN) is widely used to identify and detect gases. However, ENs are inevitably subject to sensor drift and the intrusion of unknown gases during prolonged exposure to the environment. A combined model of sensor drift compensation and open-set gas recognition, class anchoring and contrasting semantic alignment (CACSA) was proposed and tested on two commonly used EN drift datasets. Firstly, the model learns an embedding subspace that is discriminative, where mapped domains are semantically aligned and maximally separated. In addition, the method requires a very small number of labeled target training samples, even if one per class can be effective. Then, our model learns a fixed anchor for each known class. Known classes form tight clusters around the centers of class dependencies anchored in the feature space. CACSA can be degenerated into a closed-set drift compensation model without sacrificing accuracy and not require the additional target domain sample labeling and repeated training for open-set recognition. According to the experimental result, our model showed better drift compensation performance, and achieved the highest open-set recognition accuracy.

关键词： Open-set recognition Sensor drift Gas recognition Electronic nose

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：