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

arXiv: https://arxiv.org/abs/2009.04465 · HTML: https://mlsanity.com/paper/2009.04465/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 Sep 2020.

## Built on by (most cited first, 8 of 8)

- [Machine Learning for Microcontroller-Class Hardware: A Review](https://mlsanity.com/paper/2205.14550/) (2022)
- [Parallelizing Legendre Memory Unit Training](https://mlsanity.com/paper/2102.11417/) (2021)
- [Mamba-PTQ: Outlier Channels in Recurrent Large Language Models](https://mlsanity.com/paper/2407.12397/) (2024)
- [Q-S5: Towards Quantized State Space Models](https://mlsanity.com/paper/2406.09477/) (2024)
- [Efficient Neural Networks for Tiny Machine Learning: A Comprehensive Review](https://mlsanity.com/paper/2311.11883/) (2023)
- [End-to-End Efficiency in Keyword Spotting: A System-Level Approach for Embedded Microcontrollers](https://mlsanity.com/paper/2509.07051/) (2025)
- [Sub 8-Bit Quantization of Streaming Keyword Spotting Models for Embedded Chipsets](https://mlsanity.com/paper/2207.06920/) (2022)
- [QS4D: Quantization-aware training for efficient hardware deployment of structured state-space sequential models](https://mlsanity.com/paper/2507.06079/) (2025)

## Abstract (arXiv)

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 (212μW). We also characterize the power requirements of custom designed accelerator hardware that achieves SotA power efficiency of 8.79μW, beating general purpose low power hardware (a microcontroller) by 24x and special purpose ASICs by 16x.