# Metric Learning for Keyword Spotting

arXiv: https://arxiv.org/abs/2005.08776 · HTML: https://mlsanity.com/paper/2005.08776/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 May 2020.

## Built on by (most cited first, 9 of 9)

- [Deep Spoken Keyword Spotting: An Overview](https://mlsanity.com/paper/2111.10592/) (2021)
- [Meta Learning for Natural Language Processing: A Survey](https://mlsanity.com/paper/2205.01500/) (2022)
- [Learning Efficient Representations for Keyword Spotting with Triplet Loss](https://mlsanity.com/paper/2101.04792/) (2021)
- [Few-Shot Open-Set Learning for On-Device Customization of KeyWord Spotting Systems](https://mlsanity.com/paper/2306.02161/) (2023)
- [Self-Learning for Personalized Keyword Spotting on Ultra-Low-Power Audio Sensors](https://mlsanity.com/paper/2408.12481/) (2024)
- [Text Anchor Based Metric Learning for Small-footprint Keyword Spotting](https://mlsanity.com/paper/2108.05516/) (2021)
- [Encoder-Decoder Neural Architecture Optimization for Keyword Spotting](https://mlsanity.com/paper/2106.02738/) (2021)
- [Learning Decoupling Features Through Orthogonality Regularization](https://mlsanity.com/paper/2203.16772/) (2022)
- [A Quantum Kernel Learning Approach to Acoustic Modeling for Spoken Command Recognition](https://mlsanity.com/paper/2211.01263/) (2022)

## Abstract (arXiv)

The goal of this work is to train effective representations for keyword spotting via metric learning. Most existing works address keyword spotting as a closed-set classification problem, where both target and non-target keywords are predefined. Therefore, prevailing classifier-based keyword spotting systems perform poorly on non-target sounds which are unseen during the training stage, causing high false alarm rates in real-world scenarios. In reality, keyword spotting is a detection problem where predefined target keywords are detected from a variety of unknown sounds. This shares many similarities to metric learning problems in that the unseen and unknown non-target sounds must be clearly differentiated from the target keywords. However, a key difference is that the target keywords are known and predefined. To this end, we propose a new method based on metric learning that maximises the distance between target and non-target keywords, but also learns per-class weights for target keywords `a la classification objectives. Experiments on the Google Speech Commands dataset show that our method significantly reduces false alarms to unseen non-target keywords, while maintaining the overall classification accuracy.