arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1902.05026cs.LGstat.ML
收藏

一种优化的超低功耗关键字识别递归单元

An Optimized Recurrent Unit for Ultra-Low-Power Keyword Spotting

Thayer School of Engineering(塞耶工程学院);Dartmouth College(达特茅斯学院)

Justice Amoh, Kofi Odame

摘要:

人们对能够在传感器、可穿戴设备和物联网(IoT)设备上运行神经网络的兴趣越来越大。然而,神经网络的计算需求使得它们难以部署在资源受限的边缘设备上。为了满足这一需求,我们的工作介绍了一种新的经常性的单元架构,是专门适用于设备上的低功耗声事件检测(AED)。所提出的架构基于门控递归单元(GRU),但具有优化功能,使其可在超低功耗微控制器(如Arm Cortex M0+)上实现。我们的新架构,嵌入式门控递归单元(eGRU)被证明是高效的,适用于短期AED和关键字定位任务。单个eGRU单元比GRU单元快60倍,小10倍。尽管进行了优化,但eGRU在不同复杂度的任务中与GRU相比表现良好。在可穿戴声学事件检测应用中研究eGRU的实用性。在基于ARM Cortex M0的Atmel ATSAMD21E18处理器上实现并测试了eGRU模型。Arm M0+实现的eGRU模型与在工作站上运行的全精度GRU相比毫不逊色。嵌入式eGRU模型的分类准确率达到95.3%,仅比全精度GRU低2%。

英文摘要:

There is growing interest in being able to run neural networks on sensors, wearables and internet-of-things (IoT) devices. However, the computational demands of neural networks make them difficult to deploy on resource-constrained edge devices. To meet this need, our work introduces a new recurrent unit architecture that is specifically adapted for on-device low power acoustic event detection (AED). The proposed architecture is based on the gated recurrent unit (`GRU') but features optimizations that make it implementable on ultra-low power micro-controllers such as the Arm Cortex M0+. Our new architecture, the Embedded Gated Recurrent Unit (eGRU) is demonstrated to be highly efficient and suitable for short-duration AED and keyword spotting tasks. A single eGRU cell is 60x faster and 10x smaller than a GRU cell. Despite its optimizations, eGRU compares well with GRU across tasks of varying complexities. The practicality of eGRU is investigated in a wearable acoustic event detection application. An eGRU model is implemented and tested on the Arm Cortex M0-based Atmel ATSAMD21E18 processor. The Arm M0+ implementation of the eGRU model compares favorably with a full precision GRU that is running on a workstation. The embedded eGRU model achieves a classification accuracy 95.3%, which is only 2% less than the full precision GRU.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑