aiqbench analysis
This paper applies continual learning to recurrent networks for decoding hand gestures from muscle signals, comparing experience replay, elastic weight consolidation, and naive fine-tuning on NinaPro DB1. Replay retains the most accuracy under cross-user and new-gesture drift, while the authors also find that temporally ordered training can collapse recurrent models. Readers building adaptive sequence models should note that training-regime choices, not just the CL algorithm, determine whether the model learns at all.
Cross-User Adaptation for sEMG Gesture Recognition via Continual Learning
arXiv:2610.34011v1 Announce Type: new Abstract: Prosthetic and assistive interfaces increasingly depend on decoding hand gestures from surface electromyography (sEMG). Because the signal is shaped by person-specific factors, the recurrent neural networks (RNNs) commonly used to model these temporal streams must adapt to new wearers and new gesture classes after deployment. Continual learning (CL) enables this adaptation without retraining from scratch, yet applying it to recurrent sEMG models is not straightforward. Standard recurrent training practices can destabilize learning, and which CL strategy holds up under realistic cross-user drift is unestablished. We evaluate three representative CL strategies under domain- and class-incremental drift on NinaPro DB1, and find experience replay to outperform elastic weight consolidation (EWC) and naive fine-tuning. It reaches 83.3% final accuracy for class-incremental drift while limiting forgetting to 1.8-5.8% accuracy drop, as against 16.2-28.5% for EWC and naive training. We further show these models must be trained on shuffled, independently-windowed segments. Streaming windows in temporal order collapses the network into predicting a single dominant gesture, while a shuffled, stateless regime trains stably. Finally, hand-crafted time-domain features raise per-task accuracy by up to about 40% over raw signals, but only replay preserves that gain under class-incremental drift.
Abstract:Prosthetic and assistive interfaces increasingly depend on decoding hand gestures from surface electromyography (sEMG). Because the signal is shaped by person-specific factors, the recurrent neural networks (RNNs) commonly used to model these temporal streams must adapt to new wearers and new gesture classes after deployment. Continual learning (CL) enables this adaptation without retraining from scratch, yet applying it to recurrent sEMG models is not straightforward. Standard recurrent training practices can destabilize learning, and which CL strategy holds up under realistic cross-user drift is unestablished. We evaluate three representative CL strategies under domain- and class-incremental drift on NinaPro DB1, and find experience replay to outperform elastic weight consolidation (EWC) and naive fine-tuning. It reaches 83.3% final accuracy for class-incremental drift while limiting forgetting to 1.8-5.8% accuracy drop, as against 16.2-28.5% for EWC and naive training. We further show these models must be trained on shuffled, independently-windowed segments. Streaming windows in temporal order collapses the network into predicting a single dominant gesture, while a shuffled, stateless regime trains stably. Finally, hand-crafted time-domain features raise per-task accuracy by up to about 40% over raw signals, but only replay preserves that gain under class-incremental drift.
4 pages, 2 figures, 3 tables. Accepted at IEEE BioCAS 2026
Machine Learning (cs.LG); Signal Processing (eess.SP)
arXiv:2610.34011 [cs.LG]
(or arXiv:2610.34011v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.34011 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 299 words
Related coverage
Adaptive Calibration for Fair and Performant Facial Recognition
arXiv · Jun 4, 2026
De-biasing Skeleton-based Action Recognition with Convex Hull Adaptive Shift
arXiv · Sep 30, 2026
Altitude-Adaptive Vision-Only Geo-Localization for UAVs in GPS-Denied Environments
arXiv · Aug 4, 2026
Is High Local Accuracy Enough? Geometric Consensus with Support-Adaptive Statistical Personalization for Federated Learning
arXiv · Oct 23, 2026