arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2009.04465eess.AScs.LGcs.SDeess.SPstat.ML
收藏

硬件感知培训,用于在通用和专用硬件上高效识别关键字

Hardware Aware Training for Efficient Keyword Spotting on General Purpose and Specialized Hardware

Applied Brain Research Inc.(应用脑研究公司);University of Waterloo(滑铁卢大学)

Peter Blouw, Gurshaant Malik, Benjamin Morcos, Aaron R. Voelker, Chris Eliasmith

摘要:

关键字识别(KWS)为许多移动和边缘应用程序(包括手机、可穿戴设备和汽车)提供了关键的用户界面。由于KWS系统通常是“始终开启”的,因此最大限度地提高精度和功率效率是其实用性的核心。在这项工作中,我们使用硬件感知训练(HAT)来构建基于Legendre Memory Unit(LMU)的新KWS神经网络,以实现最先进的(SotA)精度和低参数计数。这允许神经网络在标准硬件(212 $μ$ W)上高效地运行。我们还对定制设计的加速器硬件的功耗要求进行了表征,该硬件实现了8.79 $μ$ W的SotA功耗效率,比通用低功耗硬件(微控制器)高24倍,比专用ASIC高16倍。

英文摘要:

Keyword spotting (KWS) provides a critical user interface for many mobile and edge applications, including phones, wearables, and cars. As KWS systems are typically 'always on', maximizing both accuracy and power efficiency are central to their utility. In this work we use hardware aware training (HAT) to build new KWS neural networks based on the Legendre Memory Unit (LMU) that achieve state-of-the-art (SotA) accuracy and low parameter counts. This allows the neural network to run efficiently on standard hardware (212$μ$W). We also characterize the power requirements of custom designed accelerator hardware that achieves SotA power efficiency of 8.79$μ$W, beating general purpose low power hardware (a microcontroller) by 24x and special purpose ASICs by 16x.

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。

↑