# On-Device Domain Learning for Keyword Spotting on Low-Power Extreme Edge Embedded Systems

arXiv: https://arxiv.org/abs/2403.10549 · HTML: https://mlsanity.com/paper/2403.10549/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 28 references that are in the map.
- First posted on arXiv 12 Mar 2024.

## Builds on (most cited first, 12 of 12)

- [Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition](https://mlsanity.com/paper/1804.03209/) (2018)
- [TensorFlow Lite Micro: Embedded Machine Learning on TinyML Systems](https://mlsanity.com/paper/2010.08678/) (2020)
- [On-Device Training Under 256KB Memory](https://mlsanity.com/paper/2206.15472/) (2022)
- [MLPerf Tiny Benchmark](https://mlsanity.com/paper/2106.07597/) (2021)
- [Hello Edge: Keyword Spotting on Microcontrollers](https://mlsanity.com/paper/1711.07128/) (2017)
- [TinyOL: TinyML with Online-Learning on Microcontrollers](https://mlsanity.com/paper/2103.08295/) (2021)
- [DORY: Automatic End-to-End Deployment of Real-World DNNs on Low-Cost IoT MCUs](https://mlsanity.com/paper/2008.07127/) (2020)
- [Deep Spoken Keyword Spotting: An Overview](https://mlsanity.com/paper/2111.10592/) (2021)
- [ConvMixer: Feature Interactive Convolution with Curriculum Learning for Small Footprint and Noisy Far-field Keyword Spotting](https://mlsanity.com/paper/2201.05863/) (2022)
- [Multi-Task Network for Noise-Robust Keyword Spotting and Speaker Verification using CTC-based Soft VAD and Global Query Attention](https://mlsanity.com/paper/2005.03867/) (2020)
- [Accelerating RNN-based Speech Enhancement on a Multi-Core MCU with Mixed FP16-INT8 Post-Training Quantization](https://mlsanity.com/paper/2210.07692/) (2022)
- [I2CR: Improving Noise Robustness on Keyword Spotting Using Inter-Intra Contrastive Regularization](https://mlsanity.com/paper/2209.06360/) (2022)

## Built on by (most cited first, 13 of 13)

- [Self-Learning for Personalized Keyword Spotting on Ultra-Low-Power Audio Sensors](https://mlsanity.com/paper/2408.12481/) (2024)
- [Neural Network Quantization for Microcontrollers: A Comprehensive Survey of Methods, Platforms, and Applications](https://mlsanity.com/paper/2508.15008/) (2025)
- [Towards smart and adaptive agents for active sensing on edge devices](https://mlsanity.com/paper/2501.06262/) (2025)
- [On-device Learning of EEGNet-based Network For Wearable Motor Imagery Brain-Computer Interface](https://mlsanity.com/paper/2409.00083/) (2024)
- [End-to-End Efficiency in Keyword Spotting: A System-Level Approach for Embedded Microcontrollers](https://mlsanity.com/paper/2509.07051/) (2025)
- [TinyML for Speech Recognition](https://mlsanity.com/paper/2504.16213/) (2025)
- [Training Multi-Layer Binary Neural Networks With Local Binary Error Signals](https://mlsanity.com/paper/2412.00119/) (2024)
- [Adaptive Noise Resilient Keyword Spotting Using One-Shot Learning](https://mlsanity.com/paper/2505.09304/) (2025)
- [Implementing Keyword Spotting on the MCUX947 Microcontroller with Integrated NPU](https://mlsanity.com/paper/2506.08911/) (2025)
- [EmbBERT: Attention Under 2 MB Memory](https://mlsanity.com/paper/2502.10001/) (2025)
- [Domain-Incremental Continual Learning for Robust and Efficient Keyword Spotting in Resource Constrained Systems](https://mlsanity.com/paper/2601.16158/) (2026)
- [OnDA: On-device Channel Pruning for Efficient Personalized Keyword Spotting](https://mlsanity.com/paper/2603.02247/) (2026)
- [What changes after deployment? A survey on On-device Learning in TinyML](https://mlsanity.com/paper/2605.31226/) (2026)

## Abstract (arXiv)

Keyword spotting accuracy degrades when neural networks are exposed to noisy environments. On-site adaptation to previously unseen noise is crucial to recovering accuracy loss, and on-device learning is required to ensure that the adaptation process happens entirely on the edge device. In this work, we propose a fully on-device domain adaptation system achieving up to 14% accuracy gains over already-robust keyword spotting models. We enable on-device learning with less than 10 kB of memory, using only 100 labeled utterances to recover 5% accuracy after adapting to the complex speech noise. We demonstrate that domain adaptation can be achieved on ultra-low-power microcontrollers with as little as 806 mJ in only 14 s on always-on, battery-operated devices.