Online Continual Learning In Keyword Spotting For Low-resource Devices Via Pooling High-order Temporal Statistics

Abstract

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.

Stats

• citations9
• S2 citations—
• github stars0
• HF likes0
• heat score7.50
• arxiv keymichieli2023online

Related papers

• Separable Temporal Convolution Plus Temporally Pooled Attention For Lightweight High-performance Keyword Spotting (2021)0.00
• Exploring Representation Learning For Small-footprint Keyword Spotting (2023)3.58
• Ctc-aligned Audio-text Embedding For Streaming Open-vocabulary Keyword Spotting (2024)3.58
• Query-by-example Keyword Spotting Using Spectral-temporal Graph Attentive Pooling And Multi-task Learning (2024)0.00
• Temporal Convolution For Real-time Keyword Spotting On Mobile Devices (2019)15.67
• GE2E-KWS: Generalized End-to-end Training And Evaluation For Zero-shot Keyword Spotting (2024)2.26
• Small-footprint Keyword Spotting Using Deep Neural Network And Connectionist Temporal Classifier (2017)0.00
• Streaming Keyword Spotting Boosted By Cross-layer Discrimination Consistency (2024)6.34