Transformer Model Boosts Activity Recognition Accuracy but Increases Computing Latency
A study evaluating the TransAct attention architecture shows accuracy and macro-F1 advantages over recurrent baselines on wearable activity data alongside ...
A study evaluating the TransAct attention architecture shows accuracy and macro-F1 advantages over recurrent baselines on wearable activity data alongside ...
Researchers developed LogISEL, an interpretable stacked machine learning framework that forecasts slope movement classes ten minutes ahead across diverse Himalayan ...
A simulation study shows that integrating PPP-RTK satellite navigation with recursive Butterworth filtering reduces UAV airborne gravimetry disturbance errors to ...
A study of 356 adults found that reciprocal associations between vigorous physical activity and muscle markers exist in young adults ...
A randomized exploratory study of 40 healthy young adults found that a 40-minute working-memory task increased perceived mental fatigue but ...
Researchers at the University of Isfahan have unveiled a wearable glove system and an ultra-compact transformer that recognize continuous multi-digit ...
© 2026 Scienmag - Science Magazine
© 2026 Scienmag - Science Magazine