Hardware Aware Training for Efficient Keyword Spotting on General Purpose and Specialized Hardware
This work addresses the need for efficient and accurate KWS for mobile and edge applications, representing a strong specific gain rather than a broad paradigm shift.
The paper tackled the problem of improving accuracy and power efficiency for keyword spotting (KWS) systems, which are always-on in mobile and edge devices, by using hardware aware training (HAT) with Legendre Memory Unit (LMU) neural networks, achieving state-of-the-art accuracy and low power consumption of 212μW on standard hardware and 8.79μW on custom accelerators.
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.