检索结果-内蒙古大学图书馆

On the Distribution of Clique-Based Neural Networks for Edge AI

ieee JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND systems 2020年第4期10卷 469-477页

作者： Larras, Benoit Frappe, Antoine Univ Lille CNRS Cent LilleUMR 8520 IEMN Univ Polytech Hauts de FranceYncrea Hauts de Fra F-59000 Lille France

distributed smart sensors are more and more used in applications such as biomedical or domestic monitoring. However, each sensor broadcasts data wirelessly to the others or to an aggregator, which leads to energy-hungry sensor nodes not ensuring data privacy. To tackle both challenges, this work proposes to distribute the feature extraction and a part of a clique-based neural network (CBNN) in each sensor node. This scheme allows standardizing data at the sensor level, ensuring privacy if the data is intercepted. Besides, a lower number of bits is transmitted, thus limiting the communication overhead. The inherent redundancy of clique-based networks makes them resilient to out-of-range connections, allowing an additional power reduction in the sensor nodes. Compared with a localized CBNN in the aggregator, the distributed structure reduces the inference latency by 28%, the sensor energy consumption by 25% and increases the protocol robustness. The circuit implementation is possible with the use of single-cluster iterative clique-based circuits, and demonstrated for a posture recognition application. To this end, a hardware circuit has been fabricated and performs a classification using 115fJ per synaptic event per neuron in 83ns.

关键词： Neural networks Wireless sensor networks Intelligent sensors Energy consumption Monitoring Feature extraction Neural networks circuit clique-based neural networks analog mixed-signal circuit distributed architecture

来源：评论

学校读者我要写书评

暂无评论

Short Course: High Speed/High Performance Data Converters: Metrics, Architectures, and Emerging Topics

Short Course: High Speed/High Performance Data Converters: M...

引用

ieee international conference on Solid-State Circuits (ISSCC)

作者： Daniel Friedman

ISBN: (数字)9781665428002

ISBN: (纸本)9781665428019

Data converters provide the gateway between the analog and digital worlds. They play a critical role in a vast range of systems, and integrated converters represent a central and growing element in high performance designs in applications from wireline to wireless and more. In this course, we start with an introduction to data converters, discussing key topologies, metrics, and the associated trade space. We continue with a discussion of ultra-high data rate converter design approaches before moving to high precision and low power converter topics. In our final session, we discuss emerging data converter concepts that will shape the future of research in this area.

关键词： Tutorials Radio frequency Wireless communication Transceivers Solid state circuits Quantum computing Optical transmitters Computer architecture Wireless sensor networks Training

来源：评论

学校读者我要写书评

暂无评论

Optimizing Traffic Controllers along the MLK Smart Corridor Using Reinforcement Learning and Digital Twin

Optimizing Traffic Controllers along the MLK Smart Corridor ...

引用

Digital Twins and Parallel Intelligence (DTPI), ieee international conference on

作者： Abhilasha Saroj Toan V. Trant Angshuman Guin Michael Hunter Mina Sartipi School of Civil and Environmental Engineering Georgia Institute of Technology GA USA Center for Urban Informatics and Progress University of Tennessee at Chattanooga TN USA

ISBN: (数字)9781665492270

ISBN: (纸本)9781665492287

With advancements in Intelligent Transportation systems (ITS), sensors, and computing resources, several cities across the world are investing in the development of smart/connected corridors. These corridors are being equipped with advanced sensors that provide real-time, high-resolution data from the corridor and enable vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V) communications. The objective of this study is to optimize signal timings for one such smart corridor – MLK Smart Corridor – in Chattanooga, Tennessee, USA with respect to fuel and energy consumption (represented by Fuel Consumption Intersection Control Performance Index, EcoPI, that determines the excess fuel consumption due to stops and delays caused by traffic controllers).

关键词： Vehicle-to-infrastructure Urban areas Vehicular ad hoc networks Reinforcement learning sensor systems Real-time systems Digital twins

来源：评论

学校读者我要写书评

暂无评论

A Scalable distributed System for Precision Irrigation 6

A Scalable Distributed System for Precision Irrigation

引用

6th ieee international conference on Smart computing (SMARTCOMP)

作者： Amoretti, Michele Rizzini, Dario Lodi Penzotti, Gabriele Caselli, Stefano Univ Parma CIDEA Parma Italy

ISBN: (纸本)9781728169972

The level of adoption of Precision Agriculture (PA) technologies is still very different from one country to another and from one region to another in the same country. A major challenge is to develop PA from a best practice pursued by a minority of enlightened farmers to a widespread practice with sizeable impact on the use of environmental resources. One of the obstacles hindering PA is the lack of quantitative data readily accessible to farmers to guide their daily operations. In this paper, we present the information system developed within project POSITIVE to support and enhance precision irrigation across the whole Emilia-Romagna region. To this purpose, the POSITIVE information system establishes a service transforming satellite and sensor data into biophysical parameters and vegetation indices with full regional coverage. These data are automatically fed into a public irrigation advisory service (Irrinet+) thereby enabling precision irrigation and fertigation on a regional scale. Irrigation maps can be sent as advice to farmers or directly commanded to registered irrigation machines. The architecture of the distributed information system and the open protocols developed to achieve scalability and enable interaction of multiple heterogeneous components are reported in the paper.

关键词： precision agriculture scalable protocols smart services

来源：评论

学校读者我要写书评

暂无评论

Pipelining Graph Construction and Agent-based Computation over distributed Memory

Pipelining Graph Construction and Agent-based Computation ov...

引用

ieee international conference on Big Data

作者： Yan Hong Munehiro Fukuda Computing and Software Systems University of Washington Bothell Bothell

ISBN: (纸本)9781665480468

Graph streaming has received substantial attention for the past 10+ years to cope with large-scale graph computation. Two major approaches, one using conventional data-streaming tools and the other accessing graph databases, facilitate continuous analysis of endlessly flowing graphs and query-based incremental construction of huge graphs, respectively. However, some scientific graphs including biological networks need to stay in memory for repetitive but various analyses. Although a cluster system, thus distributed memory can entirely handle a big graph in memory, a challenge is substantial overhead incurred by loading graphs into memory. A solution is hiding such graph-loading and construction overheads with graph computation in a pipelined fashion. We adapted this pipelining approach for agent-based graph computing where thousands of agents traverse a graph for finding its attributes and shape. We used the multi-agent spatial simulation (MASS) library to implement the concept. A huge graph is incrementally constructed in batches, each spawning and walking agents over the corresponding subgraph, and thus all eventually completing a given computation. We coded and ran two MASS benchmark programs: triangle counting and connected components, with which we evaluated our pipelined graph processing. The best performance was obtained once the batch size shrunk enough to fit cache memory, regardless of the number of cluster nodes. For a single node execution of connected components over a 140MB graph, our graph-pipelining implementation performed 7.7 times faster than non-pipelining execution. Its parallel execution with 24 cluster nodes achieved 8.3 times speed-up as compared to the pipelined single-node execution.

关键词： Shape Scalability Loading XML Big Data Size measurement Libraries

来源：评论

学校读者我要写书评

暂无评论

Charge Me Securely: Decentralized Privacy-Aware and Publicly Verifiable Energy Trading with Electric Vehicles

Charge Me Securely: Decentralized Privacy-Aware and Publicly...

引用

ieee Communications Society conference on sensor and Ad Hoc Communications and Networks (SECON)

作者： Jucai Yang Haiqin Wu Xiao Chen Zhenfu Cao Xiaolei Dong East China Normal University Shanghai China University of Leicester Leicester United Kingdom

ISBN: (数字)9798331519186

ISBN: (纸本)9798331519193

With the widespread adoption of electric vehicles (EVs), Vehicle-to-Vehicle (V2V) charging technology offers a more flexible solution for EV charging, significantly alleviating charging inconvenience, particularly in remote areas. However, the current V2V energy trading landscape lacks reliable management platforms with transparent transaction protocols. Furthermore, notable deficiencies exist in data privacy and security, which hinder the broader implementation of V2V charging services. Addressing these challenges, particularly the provision of secure and publicly auditable V2V charging without relying on centralized platforms or disclosing user privacy, has become a critical concern. For this reason, this paper presents PET, a decentralized Privacy-preserving Energy Trading system with public verifiability for EVs. PET is built upon the emerging blockchain technology to decentralize energy trading while ensuring accountability. We model energy demand-supply matching as range matching of locations and charge amounts between a buyer and sellers, and employ reverse auction to select the winner. For efficient and privacy-aware range proofs, we propose a novel batched hash chain-based range proof (BHW) primitive. In addition, PET integrates zk-SNARKs to verify payment correctness while maintaining user privacy. Our system supports public verifiability, entitling any third party to independently verify the transaction integrity. We analyze the privacy guarantees and public verifiability of PET. Extensive experiments implemented on Hyperledger Fabric further validate that PET delivers robust performance with a 4 × reduction in verification cost compared to that without batch proofs.

关键词： Privacy Costs distributed ledger Vehicular ad hoc networks Fabrics Blockchains systems support sensors Security Reliability

来源：评论

学校读者我要写书评

暂无评论

A Many-Core Accelerator Design for On-Chip Deep Reinforcement Learning 20

A Many-Core Accelerator Design for On-Chip Deep Reinforcemen...

引用

39th ieee/ACM international conference On Computer Aided Design (ICCAD)

作者： Wang, Ying Wang, Mengdi Li, Bing Li, Huawei Li, Xiaowei Chinese Acad Sci Inst Comp Technol State Key Lab Comp Architecture Beijing Peoples R China Capital Normal Univ Beijing Peoples R China

ISBN: (纸本)9781665423243

Deep Reinforcement Learning (DRL) is substantially resource-consuming, and it requires large-scale distributed computing-nodes to learn complicated tasks, like video-game and Go play. This work attempts to down-scale a distributed DRL system into a specialized many-core chip and achieve energy-efficient on-chip DRL. With the customized Network-on-Chip that handles the communication of on-chip data and control-signals, we proposed a Synchronous Asynchronous RL Architecture (SARLA) and the according many-core chip that completely avoids the unnecessary data duplication and synchronization activities in multi-node RL systems. In evaluation, the SARLA system achieves considerable energy-efficiency boost over the GPU-based implementations for typical DRL workloads built with OpenAI-gym.

关键词： reinforcement learning distributed learning many-core chip network-on-chip

来源：评论

学校读者我要写书评

暂无评论

Accurate Horse Gait Event Estimation Using an Inertial sensor Mounted on Different Body Locations

Accurate Horse Gait Event Estimation Using an Inertial Senso...

引用

international conference on Smart computing (SMARTCOMP)

作者： Hamed Darbandi Filipe Serra Braganç a Berend Jan van der Zwaag Paul Havinga Pervasive Systems Group University of Twente Enschede The Netherlands Dept. of Clinical Sciences Utrecht University Utrecht The Netherlands

Accurate calculation of temporal stride parameters is essential in horse gait analysis. A prerequisite for calculating these parameters is identifying the exact timings of gait events, i.e., hoof-on and hoof-off moments. A hoof-mounted inertial measurement unit (IMU) can be used to identify these moments accurately, yet this approach is often impractical due to the vulnerability of IMU to the impacts during locomotion. In this study, we investigated the possibility of accurately estimating the gait events using the signals of an IMU mounted on a less vulnerable location, such as a limb or upper body. To achieve the goal, we equipped IMUs on horses limbs, withers, and sacrum and measured them during different gaits. Then, we estimated the gait events timings by training recurrent neural networks models on the output signals of each IMU. Finally, we evaluated the models by comparing their results to the gait events timings labeled from hoof-mounted IMUs. The best performing model represented the best location (between the limbs, withers, and sacrum) for gait event estimation. Compared to the previous studies, our models yielded higher accuracy and were more generic by supporting more gaits. In conclusion, accurate calculation of temporal stride parameters is feasible by estimating gait event timings using an IMU mounted on less vulnerable body locations.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A distributed Approach for 5G Computation Offloading based on Non-Preemptive Priority Scheduling and Deep Reinforcement Learning

A Distributed Approach for 5G Computation Offloading based o...

引用

international conference on Computer and Communications (ICCC)

作者： Nian Liu Tao Zhang Junlang Cui Yanyi Cao Chuanfu Zhang School of Systems Science and Engineering Sun Yat-sen University Guangzhou China

Mobile devices generate numerous non-divisible and delay-sensitive tasks in 5G system, which bring a huge amount of network traffic. As Mobile edge computing can effectively alleviate computational pressure of terminal devices by offloading some tasks to edge servers, it is of great value to investigate an appropriate computation offloading scheme to reduce the sum processing delay and the task dropped rate in the 5G system. In order to deal with the problem, by adopting long short-term memory (LSTM), dueling double DQN (D3QN) and DQNReg, the problem is formulated into a long-term optimization problem in this paper. Based on non-preemptive priority scheduling and deep reinforcement learning (DQN), an innovatively distributed offloading algorithm named NPPS-D3QNReg is proposed, which can be used for adjusting the execution order of the tasks according to their deadlines. Finally, simulation results demonstrate that our proposed algorithm can effectively reduce the sum delay and task dropped rate.

关键词： Deep learning 5G mobile communication Processor scheduling Simulation Reinforcement learning Telecommunication traffic Scheduling

来源：评论

学校读者我要写书评

暂无评论

Embedded Wireless Stingless Beehive Monitoring And Data Management System

Embedded Wireless Stingless Beehive Monitoring And Data Mana...

引用

Power Engineering Application (ICPEA), ieee international conference in

作者： Noor Hafizah Khairul Anuar Mohd Amri Md Yunus Muhammad Ariff Baharudin Sallehuddin Ibrahim Shafishuhaza Sahlan Faculty of Electrical Engineering Universiti Teknologi MARA Pasir Gudang Malaysia Division School of Electrical Engineering Frontier Materials Research Alliance Control and Mechatronics Engineering Universiti Teknologi Malaysia Skudai Malaysia Control and Mechatronics Engineering School of Electrical Engineering Universiti Teknologi Malaysia Skudai Malaysia

The following topics are dealt with: photovoltaic power systems; distributed power generation; power supply quality; invertors; optimisation; PWM invertors; power engineering computing; distribution networks; neural n... 详细信息

关键词： Temperature sensors Wireless sensor networks Temperature distribution Databases System performance Humidity Strain measurement

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：