SHF: Small: Dynamic Gating and Adaptation of Deep Neural Networks for Efficient Inference and Training

SHF：小型：深度神经网络的动态门控和适应，用于高效推理和训练

• 批准号: 2007832
• 负责人: Gookwon Suh
• 金额: $ 50万
• 依托单位: Cornell University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-06-15 至 2024-05-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2007832&HistoricalAwards=false
• 关键词: SHF; Small; Dynamic; Gating; Adaptation

The past half-decade has seen unprecedented growth in machine learning with deep neural networks (DNNs), which now represent the state-of-the-art in many AI applications. However, existing DNN models require substantial memory and computing power, which greatly limit their use in resource-constrained systems such as mobile and IoT devices. This project will develop new algorithms and hardware to significantly improve the efficiency of DNNs, and represents an important step towards enabling fast and adaptive DNN executions even in resource-limited environments. In that sense, this project has the potential to enable a wider deployment of machine learning, which will play a critical role in many aspects of the future smart society. The research project will provide research training opportunities to the students as well as new curriculum development by leveraging existing resources at Cornell, e.g., summer camps as well as an outreach programs for high-school students including women.This project aims to significantly improve the efficiency of DNNs, while maintaining high accuracy, by co-developing algorithm optimizations and an efficient hardware-accelerator architecture. While there exist many lines of work on reducing DNN execution costs, the majority of these techniques are designed primarily to improve inference and perform static optimizations that reduce computation uniformly for all inputs or only exploit a limited form of dynamic sparsity, namely zeros. This project aims to enable new performance-accuracy trade-off points for DNNs that are not possible today by exploiting general forms of dynamic sparsity that are specific to each input at run-time. More specifically, the project plans to investigate input-specific gating techniques that can remove redundant computations for both training and inference, develop dynamic quantization techniques that do not require training data, and design an efficient and unified hardware accelerator architecture that provides both real-world performance and energy improvements.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.

在过去的五年中，深度神经网络（DNN）的机器学习出现了前所未有的增长，现在它代表了许多人工智能应用的最新技术。然而，现有的DNN模型需要大量的内存和计算能力，这极大地限制了它们在资源受限系统（如移动的和IoT设备）中的使用。该项目将开发新的算法和硬件，以显着提高DNN的效率，并代表即使在资源有限的环境中也能实现快速和自适应DNN执行的重要一步。从这个意义上说，该项目有可能实现更广泛的机器学习部署，这将在未来智能社会的许多方面发挥关键作用。该研究项目将为学生提供研究培训机会，并利用康奈尔大学的现有资源开发新课程，例如，该项目旨在通过共同开发算法优化和高效的硬件加速器架构，在保持高准确性的同时，显著提高DNN的效率。虽然在降低DNN执行成本方面存在许多工作，但这些技术中的大多数主要是为了改进推理和执行静态优化，这些静态优化可以均匀地减少所有输入的计算，或者只利用有限形式的动态稀疏性，即零。该项目旨在通过利用运行时特定于每个输入的一般形式的动态稀疏性，为DNN实现新的性能-精度权衡点，这在今天是不可能的。更具体地说，该项目计划研究特定于输入的门控技术，可以消除训练和推理的冗余计算，开发不需要训练数据的动态量化技术，并设计一种高效统一的硬件加速器架构，既提供真实的-该奖项反映了NSF的法定使命，并通过使用基金会的学术价值和更广泛的影响审查标准。

项目成果

MGX: near-zero overhead memory protection for data-intensive accelerators

• DOI: 10.1145/3470496.3527418
• 发表时间: 2020-04
• 期刊: Proceedings of the 49th Annual International Symposium on Computer Architecture
• 作者: Weizhe Hua;M. Umar;Zhiru Zhang;G. Suh

SoftVN: efficient memory protection via software-provided version numbers

• DOI: 10.1145/3470496.3527378
• 发表时间: 2022-06
• 期刊: Proceedings of the 49th Annual International Symposium on Computer Architecture
• 作者: M. Umar;Weizhe Hua;Zhiru Zhang;G. Suh

BulletTrain: Accelerating Robust Neural Network Training via Boundary Example Mining

• 发表时间: 2021-09
• 期刊: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
• 作者: Weizhe Hua;Yichi Zhang;Chuan Guo;Zhiru Zhang;G. Suh

PokeBNN: A Binary Pursuit of Lightweight Accuracy

• DOI: 10.1109/cvpr52688.2022.01215
• 发表时间: 2021-11
• 期刊: 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
• 作者: Yichi Zhang;Zhiru Zhang;Lukasz Lew

FracBNN: Accurate and FPGA-Efficient Binary Neural Networks with Fractional Activations

• DOI: 10.1145/3431920.3439296
• 发表时间: 2020-12
• 期刊: The 2021 ACM/SIGDA International Symposium on Field-Programmable Gate Arrays
• 作者: Yichi Zhang;Junhao Pan;Xinheng Liu;Hongzheng Chen;Deming Chen;Zhiru Zhang