Selected work

Representative Papers

A Systematic Evaluation of Self-Supervised Learning for Label-Efficient Sleep Staging with Wearable EEG

Oct 09, 2025

To address the scarcity of labeled data and high annotation costs for wearable EEG-based sleep staging, this work presents the first systematic investigation of self-supervised learning (SSL) in this domain. We propose three evaluation paradigms and conduct comprehensive comparisons of leading SSL methods on two real-world datasets—BOAS and HOGAR. Results demonstrate that SSL models achieve >80% accuracy using only 5–10% labeled data, outperforming fully supervised baselines by approximately 10%. Crucially, these models exhibit strong generalization across diverse populations, recording environments, and low signal-to-noise ratio conditions. This study not only validates the efficacy of SSL for resource-constrained neurophysiological signal analysis but also establishes a low-labeling benchmark framework for wearable EEG sleep staging—substantially lowering barriers to clinical deployment.

0 citationsRead paper

MultiPhysio-HRC: Multimodal Physiological Signals Dataset for industrial Human-Robot Collaboration

Oct 01, 2025

To address the challenge of real-time assessment of human psychophysiological states—such as stress and cognitive load—in industrial human–robot collaboration (HRC), this work introduces MultiPhysio-HRC, the first multimodal physiological dataset specifically designed for realistic HRC scenarios. The dataset synchronously records EEG, ECG, EDA, respiration (RESP), EMG, speech, and facial action units (AUs), integrated with both virtual-reality simulations and physical disassembly tasks, and is annotated using standardized subjective scales. Its key innovation lies in the first-ever integration of physiological, audio, and visual modalities within authentic industrial HRC settings, coupled with a unified state-labeling framework. Leveraging MultiPhysio-HRC, we benchmark multiple baseline models for stress and cognitive load classification. This resource provides a high-quality, open-source foundation for affective computing, adaptive robotics, and human-aware interactive systems research.

0 citationsRead paper

Recent publications

Latest Papers

A Systematic Evaluation of Self-Supervised Learning for Label-Efficient Sleep Staging with Wearable EEG

Oct 09, 2025

To address the scarcity of labeled data and high annotation costs for wearable EEG-based sleep staging, this work presents the first systematic investigation of self-supervised learning (SSL) in this domain. We propose three evaluation paradigms and conduct comprehensive comparisons of leading SSL methods on two real-world datasets—BOAS and HOGAR. Results demonstrate that SSL models achieve >80% accuracy using only 5–10% labeled data, outperforming fully supervised baselines by approximately 10%. Crucially, these models exhibit strong generalization across diverse populations, recording environments, and low signal-to-noise ratio conditions. This study not only validates the efficacy of SSL for resource-constrained neurophysiological signal analysis but also establishes a low-labeling benchmark framework for wearable EEG sleep staging—substantially lowering barriers to clinical deployment.

0 citationsRead paper

MultiPhysio-HRC: Multimodal Physiological Signals Dataset for industrial Human-Robot Collaboration

Oct 01, 2025

To address the challenge of real-time assessment of human psychophysiological states—such as stress and cognitive load—in industrial human–robot collaboration (HRC), this work introduces MultiPhysio-HRC, the first multimodal physiological dataset specifically designed for realistic HRC scenarios. The dataset synchronously records EEG, ECG, EDA, respiration (RESP), EMG, speech, and facial action units (AUs), integrated with both virtual-reality simulations and physical disassembly tasks, and is annotated using standardized subjective scales. Its key innovation lies in the first-ever integration of physiological, audio, and visual modalities within authentic industrial HRC settings, coupled with a unified state-labeling framework. Leveraging MultiPhysio-HRC, we benchmark multiple baseline models for stress and cognitive load classification. This resource provides a high-quality, open-source foundation for affective computing, adaptive robotics, and human-aware interactive systems research.

0 citationsRead paper