检索结果-内蒙古大学图书馆

Bionic cooperative load frequency control in interconnected grids: A multi-agent deep Meta reinforcement learning approach

引用

APPLIED ENERGY 2025年 379卷

作者： Li, Jiawen Dai, Jichao Cui, Haoyang Shanghai Univ Elect Power Sch Elect & Informat Engn Shanghai 201306 Peoples R China Shanghai Jiao Tong Univ Sch Elect Informat & Elect Engn Dept Elect Engn Shanghai 200240 Peoples R China

In the interconnected power grid operating within a performance-based frequency regulation market, uncoordinated frequency control strategies and power fluctuations in interconnection lines can intensify conflicts of interest among grid operators, leading to frequent and severe frequency fluctuations. To address these challenges and enhance grid stability, the Squid-Inspired Cooperative Load Frequency Control (SC-LFC) method is proposed. This method mimics the distributed neural decision-making observed in squids, treating each unit within an area as an independent agent. In real-time applications, each unit independently collects local frequency and status information, thereby avoiding coordination failures due to inter-area communication delays or errors. To achieve efficient coordinated control across multiple objectives and regions in complex, random interconnected power grids, the Automatic Curriculum Multi-Agent Deep Meta Actor-Critic (ACMA-DMAC) algorithm is introduced. This approach employs a hybrid curriculum learning strategy, enabling gradual learning and adaptation, which enhances the robustness and efficiency of the SC-LFC strategy. Simulations based on a four-area load frequency control model of the China Southern Grid (CSG) validate the effectiveness and superior performance of the proposed method.

关键词： Load frequency control Automatic curriculum multi-agent deep meta actor critic distributed neural network Interconnected grid Regulation mileage payment

来源：评论

学校读者我要写书评

暂无评论

Optimal Partitioning of distributed neural networks for Various Communication Environments 3

Optimal Partitioning of Distributed Neural Networks for Vari...

引用

3rd International Conference on Artificial Intelligence in Information and Communication (IEEE ICAIIC)

作者： Jeong, Jonghun Yang, Hoeseok Ajou Univ Dept AI Convergence Network Suwon South Korea Ajou Univ Dept Elect & Comp Engn Suwon South Korea

ISBN: (纸本)9781728176383

Recently, it is increasingly necessary to run high-end neural network applications on top of resource-constrained embedded systems, such as wearable or Internet-of-Things devices. To cope with their high computation overheads on low-end systems, the distributed neural network approach in which multiple small neural networks separately and cooperatively operate on multiple devices has been proposed. While the computational overhead could be effectively alleviated by this approach, the existing techniques still suffer from large traffics between the devices, making it vulnerable to communication failures. This drawback hinders the application of the distributed neural network techniques to wearable devices, which may be connected with each other through unstable and low data rate communication medium like human body communication. Therefore, in this paper, we propose to improve the distributed neural network by adopting a partitioning method that can adapt to given communication environments. To validate the effectiveness of the proposed portioning technique, we compare the inference accuracies of the distributed neural networks that are partitioned differently for various communication environments.

关键词： wearable device embedded system neural network distributed neural network neural network optimization

来源：评论

学校读者我要写书评

暂无评论

Implementation of Bayesian inference in distributed neural networks 26

Implementation of Bayesian inference in distributed neural n...

引用

26th Euromicro International Conference on Parallel, distributed, and network-Based Processing (PDP)

作者： Yu, Zhaofei Hang, Tiejun Liu, Jian K. Peking Univ Sch Elect Engn & Comp Sci Beijing Peoples R China Graz Univ Technol Inst Theoret Comp Sci Graz Austria

ISBN: (纸本)9781538649756

Numerous neuroscience experiments have suggested that the cognitive process of human brain is realized as probability reasoning and further modeled as Bayesian inference. It is still unclear how Bayesian inference could be implemented by neural underpinnings in the brain. Here we present a novel Bayesian inference algorithm based on importance sampling. By distributed sampling through a deep tree structure with simple and stackable basic motifs for any given neural circuit, one can perform local inference while guaranteeing the accuracy of global inference. We show that these task-independent motifs can be used in parallel for fast inference without iteration and scale-limitation. Furthermore, experimental simulations with a small-scale neural network demonstrate that our distributed sampling-based algorithm, consisting with our theoretical analysis, can approximate Bayesian inference. Taken all together, we provide a proof-of-principle to use distributed neural networks to implement Bayesian inference, which gives a road-map for large-scale Bayesian network implementation based on spiking neural networks with computer hardwares, including neuromorphic chips.

关键词： Bayesian inference distributed neural network importance sampling neural implementation

来源：评论

学校读者我要写书评

暂无评论

Missile trajectory defense planning and data simulation based on deep learning algorithm

引用

SOFT COMPUTING 2023年 1-10页

作者： Cai, Zongjian Guangzhou Coll Technol & Business Sch Marxism Guangzhou Guangdong Peoples R China

At present, distributed research on the inference process of neural networks mainly spreads the different stages of neural network functions to multiple devices. Many devices do not perform calculations simultaneously, and due to their dependence on multiple devices, their error rate is low and resource utilization is also low. On this issue, this article proposes a design scheme for distributed neural networks. Modern warfare requires increasingly high tactical requirements for high speed, high precision, and long-range strike capabilities. On the other hand, the diversity and complexity of missile missions may lead missiles to move toward informatization and intelligence. The relevant guidance and control systems will face many challenges. The environment in which missiles fly has undergone significant changes with the expansion of their range, and attack missions have evolved from unidirectional to complementary directions. Controlling self-adjusting flight and internet orbit planning have become the primary requirements for the next generation of information missiles. Therefore, this research is aimed at determining distributed neural network algorithms for coastal atmospheric characteristics, as well as protection plans for missile trajectory attacks.

关键词： distributed neural network Control system Atmospheric characteristics Missile trajectory

来源：评论

学校读者我要写书评

暂无评论

distributed AND LOCAL neural CLASSIFIERS FOR PHONEME RECOGNITION

引用

PATTERN RECOGNITION LETTERS 1994年第11期15卷 1111-1118页

作者： GURGEN, F ALPAYDIN, R UNLUAKIN, U ALPAYDIN, E Department of Computer Engineering Boǧaziçi University TR-80815 Istanbul Turkey

The comparative performances of distributed and local neural networks for the speech recognition problem are investigated. We consider a feed-forward network with one or more hidden layers. Depending on the response characteristics of the hidden units, we name the network distributed or local. If the hidden units use the sigmoid non-linearity, then hidden units have a global response and we call such networks distributed. If each hidden unit responds only to inputs in a certain local region in the input space, then the network is local. Neighbor and prototype based approaches are of this type. As examples of the distributed approach with sigmoidal units, we employ the back-propagation rule with three error measures: mean square error, cross entropy, and combinational performance. As for the local methods, we use k-nearest neighbor, learning Vector quantization, grow and learn, and Gaussian-based weighted approximation methods. Phoneme recognition experiments are conducted using the /b, d, g, m, n, N/ set of the Japanese vocabulary for the speaker dependent case. Three criteria are taken for comparison: correct classification of the test set, network size, and learning time. We found that distributed networks generalize better than local networks but require longer training and more precise computation. Local networks learn very quickly, but do not generalize well and use more memory.

关键词： SPEECH RECOGNITION ARTIFICIAL neural networkS distributed neural network BACKPROPAGATION LOCAL networkS K-NEAREST NEIGHBOR RULE GROW AND LEARN METHOD LEARNING VECTOR QUANTIZATION

来源：评论

学校读者我要写书评

暂无评论

distributed neural Precoding for Hybrid mmWave MIMO Communications With Limited Feedback

引用

IEEE COMMUNICATIONS LETTERS 2022年第7期26卷 1568-1572页

作者： Wei, Kai Xu, Jindan Xu, Wei Wang, Ning Chen, Dong Southeast Univ Natl Mobile Commun Res Lab NCRL Nanjing 210096 Peoples R China Zhengzhou Univ Henan Joint Int Res Lab Intelligent Networking & Zhengzhou 450001 Peoples R China Zhengzhou Univ Dept Elect Engn Zhengzhou 450001 Henan Peoples R China Xiaomi Corp Beijing 100085 Peoples R China

Hybrid precoding is a cost-efficient technique for millimeter wave (mmWave) massive multiple-input multiple-output (MIMO) communications. This paper proposes a deep learning approach by using a distributed neural network for hybrid analog-and-digital precoding design with limited feedback. The proposed distributed neural precoding network, called DNet, is committed to achieving two objectives. First, the DNet realizes channel state information (CSI) compression with a distributed architecture of neural networks, which enables practical deployment on multiple users. Specifically, this neural network is composed of multiple independent sub-networks with the same structure and parameters, which reduces both the number of training parameters and network complexity. Secondly, DNet learns the calculation of hybrid precoding from reconstructed CSI from limited feedback. Different from existing black-box neural network design, the DNet is specifically designed according to the data form of the matrix calculation of hybrid precoding. Simulation results show that the proposed DNet significantly improves the performance up to nearly 50% compared to traditional limited feedback precoding methods under the tests with various CSI compression ratios.

关键词： Precoding neural networks Radio frequency Decoding Training Massive MIMO Signal to noise ratio distributed neural network hybrid precoding limited feedback millimeter wave (mmWave) MIMO

来源：评论

学校读者我要写书评

暂无评论

Excitation-inhibition balance regulates the patterning of spinal and cranial inspiratory motor outputs in rats in situ

引用

RESPIRATORY PHYSIOLOGY & NEUROBIOLOGY 2019年 266卷 95-102页

作者： Dhingra, Rishi R. Furuya, Werner I. Galan, Roberto F. Dutschmann, Mathias Florey Inst Neurosci & Mental Hlth Div Syst Neurophysiol Melbourne Vic Australia Case Western Reserve Univ Dept Elect Engn & Comp Sci Cleveland OH 44106 USA

Spinal phrenic nerve activity (PNA) drives the diaphragm but cranial hypoglossal nerve activity (HNA) also expresses synchronous activity during inspiration. Here, we investigated the effects of local disinhibition (bilateral microinjections of bicuculline) of the nucleus tractus solitarius (NTS), the pre-Botzinger complex and Botzinger complex core circuit (pre-BotC/BotC) and the Kolliker-Fuse nuclei (KFn) on the synchronization of PNA and HNA in arterially-perfused brainstem preparations of rats. To quantitatively analyze the bicuculline effects on a putatively distributed inspiratory central pattern generator (i-CPG), we quantified the phase synchronization properties between PNA and HNA. The analysis revealed that bicuculline-evoked local disinhibition significantly reduced the strength of phase synchronization between PNA and HNA at any target site. However, the emergence of desynchronized HNA following disinhibition was more prevalent after NTS or pre-BotC/BotC microinjections compared to the KFn. We conclude that the primary i-CPG is located in a distributed medullary circuit whereas pontine contributions are restricted to synaptic gating of synchronous HNA and PNA.

关键词： Respiratory pattern formation distributed neural network Excitation-inhibition balance Respiratory microcircuit

来源：评论

学校读者我要写书评

暂无评论

A novel framework to enhance the performance of training distributed deep neural networks

引用

INTELLIGENT DATA ANALYSIS 2023年第3期27卷 753-768页

作者： Phan, Trung Do, Phuc Univ Informat Technol Vietnam Natl Univ Fac Informat Sci & Engn Ho Chi Minh City Vietnam Hoa Sen Univ Fac Informat Technol Ho Chi Minh City Vietnam

There are many attempts to implement deep neural network (DNN) distributed training frameworks. In these attempts, Apache Spark was used to develop the frameworks. Each framework has its advantages and disadvantages and needs further improvements. In the process of using Apache Spark to implement distributed training systems, we ran into some obstacles that significantly affect the performance of the systems and programming thinking. This is the reason why we developed our own distributed training framework, called distributed Deep Learning Framework (DDLF), which is completely independent of Apache Spark. Our proposed framework can overcome the obstacles and is highly scalable. DDLF helps to develop applications that train DNN in a distributed environment (referred to as distributed training) in a simple, natural, and flexible way. In this paper, we will analyze the obstacles when implementing a distributed training system on Apache Spark and present solutions to overcome them in DDLF. We also present the features of DDLF and how to implement a distributed DNN training application on this framework. In addition, we conduct experiments by training a Convolutional neural network (CNN) model with datasets MNIST and CIFAR-10 in Apache Spark cluster and DDLF cluster to demonstrate the flexibility and effectiveness of DDLF.

关键词： distributed deep learning framework distributed neural network distributed processing system distributed training Apache Spark

来源：评论

学校读者我要写书评

暂无评论

Study on activated sludge wastewater treatment process intelligent control based on soft-sensing

Study on activated sludge wastewater treatment process intel...

引用

IEEE International Conference on Control and Automation

作者： Feng Yu-zhao Zhuo Ming Tian Li-qiang Zhao Hong-wei Logist Engn Univ Lab Ctr Water Supply & Environm Based Field Opera Chongqing Peoples R China

ISBN: (纸本)9781424408177

Aiming at the difficult of the control system for activated sludge wastewater treatment, the intelligent control strategy based on soft-sensing is proposed. By using the method of soft-sensing and distribute neural network respective, the soft-sensing models are built. From the point of the growth and progentiting of the microbe, discusses and completes the intelligent control soft-sensing secondary variable selecting. By analyzing the growth and propagating of the activated sludge microbe, the output of the experiment indicates that the precision of soft-sensing model is high enough to use, thus, approve the correctness. The result of imitating indicates that the practicability and validity of research way, then points out that the on-line real-time computing was the direction in the future research.

关键词： wastewater treatment distributed neural network secondary variable intelligent control soft-sensing

来源：评论

学校读者我要写书评

暂无评论

DCP-Net: A distributed Collaborative Perception network for Remote Sensing Semantic Segmentation

引用

REMOTE SENSING 2024年第13期16卷

作者： Wang, Zhechao Cheng, Peirui Duan, Shujing Chen, Kaiqiang Wang, Zhirui Li, Xinming Sun, Xian Chinese Acad Sci Aerosp Informat Res Inst Beijing 100190 Peoples R China Chinese Acad Sci Aerosp Informat Res Inst Key Lab Network Informat Syst Technol NIST Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Elect Elect & Commun Engn Beijing 100190 Peoples R China

Collaborative perception enhances onboard perceptual capability by integrating features from other platforms, effectively mitigating the compromised accuracy caused by a restricted observational range and vulnerability to interference. However, current implementations of collaborative perception overlook the prevalent issues of both limited and low-reliability communication, as well as misaligned observations in remote sensing. To address this problem, this article presents an innovative distributed collaborative perception network (DCP-Net) specifically designed for remote sensing applications. Firstly, a self-mutual information match module is proposed to identify collaboration opportunities and select suitable partners. This module prioritizes critical collaborative features and reduces redundant transmission for better adaptation to weak communication in remote sensing. Secondly, a related feature fusion module is devised to tackle the misalignment between local and collaborative features due to the multiangle observations, improving the quality of fused features for the downstream task. We conduct extensive experiments and visualization analyses using three semantic segmentation datasets, namely Potsdam, iSAID, and DFC23. The results demonstrate that DCP-Net outperforms the existing collaborative perception methods comprehensively, improving mIoU by 2.61% to 16.89% at the highest collaboration efficiency and achieving state-of-the-art performance.

关键词： collaborative perception distributed neural network semantic segmentation remote sensing

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：