# Conditional Online Learning for Keyword Spotting

arXiv: https://arxiv.org/abs/2305.13332 · HTML: https://mlsanity.com/paper/2305.13332/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 21 references that are in the map.
- First posted on arXiv 19 May 2023.

## Builds on (most cited first, 7 of 7)

- [Federated Learning with Non-IID Data](https://mlsanity.com/paper/1806.00582/) (2018)
- [Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition](https://mlsanity.com/paper/1804.03209/) (2018)
- [Unsupervised Hard Example Mining from Videos for Improved Object Detection](https://mlsanity.com/paper/1808.04285/) (2018)
- [Online Continual Learning of End-to-End Speech Recognition Models](https://mlsanity.com/paper/2207.05071/) (2022)
- [Optimize what matters: Training DNN-HMM Keyword Spotting Model Using End Metric](https://mlsanity.com/paper/2011.01151/) (2020)
- [Progressive Continual Learning for Spoken Keyword Spotting](https://mlsanity.com/paper/2201.12546/) (2022)
- [Rainbow Keywords: Efficient Incremental Learning for Online Spoken Keyword Spotting](https://mlsanity.com/paper/2203.16361/) (2022)

## Abstract (arXiv)

Modern approaches for keyword spotting rely on training deep neural networks on large static datasets with i.i.d. distributions. However, the resulting models tend to underperform when presented with changing data regimes in real-life applications. This work investigates a simple but effective online continual learning method that updates a keyword spotter on-device via SGD as new data becomes available. Contrary to previous research, this work focuses on learning the same KWS task, which covers most commercial applications. During experiments with dynamic audio streams in different scenarios, that method improves the performance of a pre-trained small-footprint model by 34%. Moreover, experiments demonstrate that, compared to a naive online learning implementation, conditional model updates based on its performance in a small hold-out set drawn from the training distribution mitigate catastrophic forgetting.