SHF: Small: Cross-Platform Solutions for Pruning and Accelerating Neural Network Models

SHF：小型：用于修剪和加速神经网络模型的跨平台解决方案

• 批准号: 1615475
• 负责人: Hai Li
• 金额: $ 45万
• 依托单位: University of Pittsburgh
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-07-01 至 2017-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1615475&HistoricalAwards=false
• 关键词: SHF; Small; Cross; Platform; Solutions

Deep neural networks (DNNs) have achieved remarkable success in many applications because of their powerful capability for data processing. The objective of this project is to investigate a software-hardware co-design methodology for DNN acceleration that can be applied to both traditional von Neumann and emerging neuromorphic architectures. The project fits into the general area of "brain-inspired" energy efficient computing paradigms that has been of much recent interest. The investigators are also active in various outreach and educational activities that include curricular development, engagement of minority/underrepresented students in research. Undergraduate and graduate students involved in this research will also be trained for the next-generation computer engineering and semiconductor industry workforce.From a more technical standpoint, a novel neural network sparsification process is to be explored to preserve the state-of-the-art accuracy, while establishing hardware-friendly models of neural network computations. The result is expected to lead to a holistic methodology composed of neural network model sparsification, hardware acceleration, and an integrated software/hardware co-design. The project also benefits big data research and industry at large by inspiring an interactive design philosophy between the design of learning algorithms and the corresponding computational platforms for system performance and scalability enhancement.

深度神经网络以其强大的数据处理能力在许多应用中取得了显著的成功。该项目的目标是研究一种用于DNN加速的软硬件协同设计方法，该方法可应用于传统的冯·诺伊曼和新兴的神经形态架构。该项目符合最近备受关注的“大脑启发”节能计算范式的一般领域。调查人员还积极参与各种外联和教育活动，包括课程开发，少数民族/代表性不足的学生参与研究。参与这项研究的本科生和研究生也将为下一代计算机工程和半导体行业的劳动力进行培训。从更技术性的角度来看，需要探索一种新的神经网络稀疏化过程，以保持最先进的精度，同时建立神经网络计算的硬件友好模型。该结果有望导致一种由神经网络模型稀疏化、硬件加速和集成软件/硬件协同设计组成的整体方法。该项目还通过激发学习算法设计与相应的计算平台之间的交互设计理念，以增强系统性能和可扩展性，从而使大数据研究和整个行业受益。