mlsanityData to 14 Oct 2026

Online Continual Learning in Keyword Spotting for Low-Resource Devices via Pooling High-Order Temporal Statistics

arXiv 2307.12660 ↗First posted 24 Jul 2023cs.SD cs.LG eess.AScited by 9

Not on the map. Why?

Abstract (arXiv)

Keyword Spotting (KWS) models on embedded devices should adapt fast to new user-defined words without forgetting previous ones. Embedded devices have limited storage and computational resources, thus, they cannot save samples or update large models. We consider the setup of embedded online continual learning (EOCL), where KWS models with frozen backbone are trained to incrementally recognize new words from a non-repeated stream of samples, seen one at a time. To this end, we propose Temporal Aware Pooling (TAP) which constructs an enriched feature space computing high-order moments of speech features extracted by a pre-trained backbone. Our method, TAP-SLDA, updates a Gaussian model for each class on the enriched feature space to effectively use audio representations. In experimental analyses, TAP-SLDA outperforms competitors on several setups, backbones, and baselines, bringing a relative average gain of 11.3% on the GSC dataset.

Key facts. Cited by 9 papers in the map (data through 14 Oct 2026). Builds on 26 of its 45 references that are in the map. First posted on arXiv 24 Jul 2023.

Builds on

26 of 45 references.

References · 26

  1. 2016Overcoming catastrophic forgetting in neural networkscited by 4,787
  2. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,905
  3. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,342
  4. 2016iCaRL: Incremental Classifier and Representation Learningcited by 2,333
  5. 2018Efficient Lifelong Learning with A-GEMcited by 1,058
  6. 2018Speech Commands: A Dataset for Limited-Vocabulary Speech Recognitioncited by 752
  7. 2018End-to-End Incremental Learningcited by 597
  8. 2021Online Continual Learning in Image Classification: An Empirical Surveycited by 245
  9. 2021Supervised Contrastive Replay: Revisiting the Nearest Class Mean Classifier in Online Class-Incremental Continual Learningcited by 141
  10. 2022FeTrIL: Feature Translation for Exemplar-Free Class-Incremental Learningcited by 121
  11. 2020Emformer: Efficient Memory Transformer Based Acoustic Model For Low Latency Streaming Speech Recognitioncited by 113
  12. 2019Lifelong Machine Learning with Deep Streaming Linear Discriminant Analysiscited by 92
14 more
  1. 2021UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Datacited by 77
  2. 2017Deep Residual Learning for Small-Footprint Keyword Spottingcited by 68
  3. 2017Towards Faster Training of Global Covariance Pooling Networks by Iterative Matrix Square Root Normalizationcited by 65
  4. 2022Online Continual Learning for Embedded Devicescited by 49
  5. 2013Stochastic Pooling for Regularization of Deep Convolutional Neural Networkscited by 48
  6. 2021Deep Spoken Keyword Spotting: An Overviewcited by 43
  7. 2021Continual Learning at the Edge: Real-Time Training on Smartphone Devicescited by 20
  8. 2022Online Continual Learning of End-to-End Speech Recognition Modelscited by 15
  9. 2021F-SIOL-310: A Robotic Dataset and Benchmark for Few-Shot Incremental Object Learningcited by 11
  10. 2022Progressive Continual Learning for Spoken Keyword Spottingcited by 8
  11. 2022Rainbow Keywords: Efficient Incremental Learning for Online Spoken Keyword Spottingcited by 8
  12. 2019Multi-Task Learning with High-Order Statistics for X-vector based Text-Independent Speaker Verificationcited by 4
  13. 2021Study on the temporal pooling used in deep neural networks for speaker verificationcited by 3
  14. 2023On the challenges to learn from Natural Data Streamscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 4; 2025: 4; 2026: 1.

9 papers.

  1. 2024Recent Advances of Foundation Language Models-based Continual Learning: A Surveycited by 46
  2. 2024Continual Learning Optimizations for Auto-regressive Decoder of Multilingual ASR systemscited by 7
  3. 2024HOP to the Next Tasks and Domains for Continual Learning in NLPcited by 3
  4. 2024Object-conditioned Bag of Instances for Few-Shot Personalized Instance Recognitioncited by 2
  5. 2025Efficient Continual Learning in Keyword Spotting using Binary Neural Networkscited by 2
  6. 2025Language-Aware Prompt Tuning for Parameter-Efficient Seamless Language Expansion in Multilingual ASRcited by 2
  7. 2025Adaptive Noise Resilient Keyword Spotting Using One-Shot Learningcited by 1
  8. 2025MetaCLBench: Meta Continual Learning Benchmark on Resource-Constrained Edge Devicescited by 0
  9. 2026Domain-Incremental Continual Learning for Robust and Efficient Keyword Spotting in Resource Constrained Systemscited by 0