SHF: Small: A General Framework for Accelerating AI on Resource-Constrained Edge Devices

SHF:小型:在资源受限的边缘设备上加速 AI 的通用框架

基本信息

  • 批准号:
    2211163
  • 负责人:
  • 金额:
    $ 60万
  • 依托单位:
  • 依托单位国家:
    美国
  • 项目类别:
    Standard Grant
  • 财政年份:
    2022
  • 资助国家:
    美国
  • 起止时间:
    2022-07-01 至 2025-06-30
  • 项目状态:
    未结题

项目摘要

The upward trend of the pervasive usage of edge devices provides excellent opportunities for on-device intelligence in future mobile and IoT applications, including mobile augmented reality (AR)/Virtual reality (VR), smart manufacturing, mobile healthcare, and autonomous vehicles. While these edge devices have complete software/hardware stacks to execute machine-learning models, they usually have constrained computing resources. They cannot afford to execute the machine-learning models directly. To keep up with the fast-growing deployment of mobile and IoT applications, it is urgently needed to design new neural-network architectures for accelerating artificial intelligence (AI) on resource-constrained edge devices. This proposal aims to develop a novel framework that can efficiently design neural-network architectures suitable for execution on edge devices. The proposed framework develops network architectures that simultaneously balance memory cost, computing efficiency, and prediction accuracy, which can advance on-device AI applications with low-latency and high-efficiency requirements. The new deployment optimization methods can generally benefit neural-network implementation and deployment on heterogeneous commodity computing platforms without customized hardware. The project will lead to a solid foundation for a broad range of research topics related to computing architecture design and edge-computing systems. The research results can benefit interdisciplinary curriculums with new research topics and tasks for undergraduate/graduate and minority students.This project develops a holistic framework for designing efficient and effective neural-network architectures considering edge devices’ hardware constraints. It first develops an automated hardware-aware neural-network architecture-design method to efficiently generate optimal neural-network architectures that can balance the trade-offs between the required accuracy and computational performance. A further investigation is conducted to develop novel neural-network optimization methods to reduce memory footprints and computational costs in a fine-grained way while satisfying the accuracy requirement. New pruning methods are designed to accurately track the importance of network parameters and effectively reduce pruning iterations and floating-point operations. Moreover, novel implementation mechanisms, such as weight-sharing-aware fine-tuning, dynamic partitioning, and on-demand loading schemes, are developed to minimize the loading time overhead and enable efficient deployment and evaluation of the designed architectures on edge devices. Commodity edge devices and FPGAs are employed to implement and evaluate the designed neural-network architecture.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
边缘设备的普遍使用的上升趋势为未来移动和物联网应用中的设备智能提供了绝佳的机会,包括移动增强现实(AR)/虚拟现实(VR)、智能制造、移动医疗和自动驾驶汽车。边缘设备具有完整的软件/硬件堆栈来执行机器学习模型,但它们通常无法直接执行机器学习模型,以跟上移动和物联网应用程序的快速增长。迫切需要设计新的用于在资源受限的边缘设备上加速人工智能(AI)的神经网络架构该提案旨在开发一种新颖的框架,可以有效地设计适合在边缘设备上执行的神经网络架构。该框架开发了同时平衡的网络架构。新的部署优化方法通常有利于神经网络在异构商品计算平台上的实现和部署,而无需定制硬件。 .该项目将引领为与计算架构设计和边缘计算系统相关的广泛研究主题奠定坚实的基础。研究成果可以使跨学科课程受益,为本科生/研究生和少数族裔学生提供新的研究主题和任务。该项目开发了一个整体框架。考虑边缘设备的限制,设计高效且有效的神经网络架构,它首先开发一种自动化的硬件感知神经网络架构设计方法,以生成有效的最佳神经网络架构,以平衡所需的精度和计算之间的权衡。正在进行进一步调查以了解性能。开发新颖的神经网络优化方法,以细粒度的方式减少内存占用和计算成本,同时满足精度要求。新的剪枝方法旨在准确跟踪网络参数的重要性,并有效减少剪枝迭代和浮点运算。此外,开发了新的实现机制,例如权重共享感知微调、动态分区和按需加载方案,以最大限度地减少加载时间开销,并能够在边缘设备上部署和评估所设计的高效架构。边缘设备和采用 FPGA 来实现和评估所设计的神经网络架构。该奖项反映了 NSF 的法定使命,并通过使用基金会的智力优势和更广泛的影响审查标准进行评估,被认为值得支持。

项目成果

期刊论文数量(3)
专著数量(0)
科研奖励数量(0)
会议论文数量(0)
专利数量(0)
Universal Targeted Adversarial Attacks Against mmWave-based Human Activity Recognition
针对基于毫米波的人类活动识别的通用定向对抗攻击
Universal Targeted Adversarial Attacks Against mmWave-based Human Activity Recognition
针对基于毫米波的人类活动识别的通用定向对抗攻击
Person re-identification using wifi signals
利用wifi信号进行人员重新识别
{{ item.title }}
{{ item.translation_title }}
  • DOI:
    {{ item.doi }}
  • 发表时间:
    {{ item.publish_year }}
  • 期刊:
  • 影响因子:
    {{ item.factor }}
  • 作者:
    {{ item.authors }}
  • 通讯作者:
    {{ item.author }}

数据更新时间:{{ journalArticles.updateTime }}

{{ item.title }}
  • 作者:
    {{ item.author }}

数据更新时间:{{ monograph.updateTime }}

{{ item.title }}
  • 作者:
    {{ item.author }}

数据更新时间:{{ sciAawards.updateTime }}

{{ item.title }}
  • 作者:
    {{ item.author }}

数据更新时间:{{ conferencePapers.updateTime }}

{{ item.title }}
  • 作者:
    {{ item.author }}

数据更新时间:{{ patent.updateTime }}

Yingying Chen其他文献

Syntheses, structures, and host-guest interactions of 2-D grid-type cyanide-bridged compounds [Zn(L)(H2O)2][M(CN)4]·3H2O (L = N,N′-bis(4-pyridylformamide)-1,4-benzene; M = Ni, Pd or Pt)
二维网格型氰化物桥化合物[Zn(L)(H2O)2][M(CN)4]·3H2O (L = N,Nâ²-bis)的合成、结构和主客体相互作用
  • DOI:
    10.1080/00958972.2013.832229
  • 发表时间:
    2013-08-07
  • 期刊:
  • 影响因子:
    1.9
  • 作者:
    Ai;Xin Chen;Hu Zhou;Yingying Chen;Aihua Yuan
  • 通讯作者:
    Aihua Yuan
The Piezo channel is central to the mechano-sensitive channel complex in the mammalian inner ear.
压电通道是哺乳动物内耳中机械敏感通道复合体的中心。
  • DOI:
    10.21203/rs.3.rs-2287052/v1
  • 发表时间:
    2023-07-12
  • 期刊:
  • 影响因子:
    0
  • 作者:
    J. Lee;Cristina M. Perez;Seojin Park;H. J. Kim;Yingying Chen;Mincheol Kang;Jennifer Kersigo;Jinsil Choi;Phung N. Thai;Ryan L Woltz;G. Perkins;Choong;Bernd Fritzsch;Pauline Trinh;Xiao;N. Chiamvimonvat;D. Perez;Padmini Sirish;Yao Dong;I. Pessah;Feng Wei;R. Dixon;B. Sokolowski;E. Yamoah
  • 通讯作者:
    E. Yamoah
Numerical Simulations of Ice Loads on Fixed and Floating Offshore Structures using the Discrete Element Method
使用离散元法对固定和浮动海上结构的冰荷载进行数值模拟
  • DOI:
    10.4043/27414-ms
  • 发表时间:
    2016-10-24
  • 期刊:
  • 影响因子:
    0
  • 作者:
    Jiancheng Liu;Xiang Liu;Yingying Chen;X. Long;S. Ji
  • 通讯作者:
    S. Ji
Effects of interfacial contact under different operating conditions in proton exchange membrane water electrolysis
质子交换膜水电解不同操作条件下界面接触的影响
  • DOI:
    10.1016/j.electacta.2022.140942
  • 发表时间:
    2022-08-01
  • 期刊:
  • 影响因子:
    6.6
  • 作者:
    Zhenye Kang;Tobias Schuler;Yingying Chen;Min Wang;Feng;G. Bender
  • 通讯作者:
    G. Bender
Adenosine monophosphate boosts the cryoprotection of ultrasound-assisted freezing to frozen surimi: Insights into protein structures and gelling behaviors.
单磷酸腺苷增强超声波辅助冷冻对冷冻鱼糜的冷冻保护:深入了解蛋白质结构和胶凝行为。
  • DOI:
    10.1016/j.foodchem.2024.139343
  • 发表时间:
    2024-04-01
  • 期刊:
  • 影响因子:
    8.8
  • 作者:
    Xiangwei Zhu;Di He;Yingying Chen;Xinyu Duan;Yonghui Li;Yue Yuan;Fuchao Zhan;Bin Li;Yongxin Teng
  • 通讯作者:
    Yongxin Teng

Yingying Chen的其他文献

{{ item.title }}
{{ item.translation_title }}
  • DOI:
    {{ item.doi }}
  • 发表时间:
    {{ item.publish_year }}
  • 期刊:
  • 影响因子:
    {{ item.factor }}
  • 作者:
    {{ item.authors }}
  • 通讯作者:
    {{ item.author }}

{{ truncateString('Yingying Chen', 18)}}的其他基金

Collaborative Research: III: Small: Efficient and Robust Multi-model Data Analytics for Edge Computing
协作研究:III:小型:边缘计算的高效、稳健的多模型数据分析
  • 批准号:
    2311596
  • 财政年份:
    2023
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
Collaborative Research: SaTC: CORE: Small: Securing IoT and Edge Devices under Audio Adversarial Attacks
协作研究:SaTC:核心:小型:在音频对抗攻击下保护物联网和边缘设备
  • 批准号:
    2114220
  • 财政年份:
    2021
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
Collaborative Research: CCRI: New: Nation-wide Community-based Mobile Edge Sensing and Computing Testbeds
合作研究:CCRI:新:全国范围内基于社区的移动边缘传感和计算测试平台
  • 批准号:
    2120396
  • 财政年份:
    2021
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
Collaborative Research: PPoSS: Planning: Hardware-accelerated Trustworthy Deep Neural Network
合作研究:PPoSS:规划:硬件加速的可信深度神经网络
  • 批准号:
    2028876
  • 财政年份:
    2020
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
SHF: Small: Collaborative Research: Software Hardware Architecture Co-design for Low-power Heterogeneous Edge Devices
SHF:小型:协作研究:低功耗异构边缘设备的软件硬件架构协同设计
  • 批准号:
    1909963
  • 财政年份:
    2019
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
SaTC: CORE: Small: Collaborative: Security Assurance in Short Range Communication with Wireless Channel Obfuscation
SaTC:核心:小型:协作:通过无线信道混淆实现短距离通信的安全保证
  • 批准号:
    1814590
  • 财政年份:
    2018
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
NeTS: Medium: Collaborative Research: Exploiting Fine-grained WiFi Signals for Wellbeing Monitoring
NeTS:媒介:协作研究:利用细粒度 WiFi 信号进行健康监测
  • 批准号:
    1826647
  • 财政年份:
    2017
  • 资助金额:
    $ 60万
  • 项目类别:
    Continuing Grant
SaTC: CORE: Small: Collaborative: Exploiting Physical Properties in Wireless Networks for Implicit Authentication
SaTC:核心:小型:协作:利用无线网络中的物理属性进行隐式身份验证
  • 批准号:
    1820624
  • 财政年份:
    2017
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
SaTC: CORE: Small: Collaborative: Exploiting Physical Properties in Wireless Networks for Implicit Authentication
SaTC:核心:小型:协作:利用无线网络中的物理属性进行隐式身份验证
  • 批准号:
    1716500
  • 财政年份:
    2017
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
NeTS: Medium: Collaborative Research: Exploiting Fine-grained WiFi Signals for Wellbeing Monitoring
NeTS:媒介:协作研究:利用细粒度 WiFi 信号进行健康监测
  • 批准号:
    1514436
  • 财政年份:
    2015
  • 资助金额:
    $ 60万
  • 项目类别:
    Continuing Grant

相似国自然基金

ALKBH5介导的SOCS3-m6A去甲基化修饰在颅脑损伤后小胶质细胞炎性激活中的调控作用及机制研究
  • 批准号:
    82301557
  • 批准年份:
    2023
  • 资助金额:
    30 万元
  • 项目类别:
    青年科学基金项目
miRNA前体小肽miPEP在葡萄低温胁迫抗性中的功能研究
  • 批准号:
  • 批准年份:
    2023
  • 资助金额:
    50 万元
  • 项目类别:
PKM2苏木化修饰调节非小细胞肺癌起始细胞介导的耐药生态位的机制研究
  • 批准号:
    82372852
  • 批准年份:
    2023
  • 资助金额:
    49 万元
  • 项目类别:
    面上项目
基于翻译组学理论探究LncRNA H19编码多肽PELRM促进小胶质细胞活化介导电针巨刺改善膝关节术后疼痛的机制研究
  • 批准号:
    82305399
  • 批准年份:
    2023
  • 资助金额:
    30 万元
  • 项目类别:
    青年科学基金项目
CLDN6高表达肿瘤细胞亚群在非小细胞肺癌ICB治疗抗性形成中的作用及机制研究
  • 批准号:
    82373364
  • 批准年份:
    2023
  • 资助金额:
    49 万元
  • 项目类别:
    面上项目

相似海外基金

SHF: Small: A General-purpose Parallel and Heterogeneous Task Graph Computing System for VLSI CAD
SHF:小型:用于 VLSI CAD 的通用并行异构任务图计算系统
  • 批准号:
    2349141
  • 财政年份:
    2023
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
Collaborative Research: SHF: Small: A General Framework for Responsive Static Analysis
合作研究:SHF:小型:响应式静态分析的通用框架
  • 批准号:
    2223825
  • 财政年份:
    2022
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
Collaborative Research: SHF: Small: A General Framework for Responsive Static Analysis
合作研究:SHF:小型:响应式静态分析的通用框架
  • 批准号:
    2223826
  • 财政年份:
    2022
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
SHF: Small: Ubiquitous and Transparent Near-data Computing for General Purpose Processors
SHF:小型:通用处理器的无处不在且透明的近数据计算
  • 批准号:
    2200831
  • 财政年份:
    2022
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
SHF: Small: A General-purpose Parallel and Heterogeneous Task Graph Computing System for VLSI CAD
SHF:小型:用于 VLSI CAD 的通用并行异构任务图计算系统
  • 批准号:
    2126672
  • 财政年份:
    2021
  • 资助金额:
    $ 60万
  • 项目类别:
    Standard Grant
{{ showInfoDetail.title }}

作者:{{ showInfoDetail.author }}

知道了