Human Heterogeneity Invariant Stress Sensing

2025-06-02

AI Summary

To address poor model generalizability in stress detection—caused by inter-individual variability, health conditions (e.g., opioid use disorder), and diverse real-world scenarios—this paper proposes a target-domain-data-free domain generalization method. We innovatively design a subject-specific subnet pruning intersection mechanism to extract cross-subject shared physiological stress features, coupled with continuous-label constraint training to mitigate overfitting. The method fuses heterogeneous multimodal physiological signals (ECG, EDA, PPG, etc.) and enables low-latency, lightweight deployment. Evaluated on seven real-world stress datasets spanning laboratory, driving, and field settings, our approach significantly outperforms existing state-of-the-art methods. Results demonstrate strong robustness, superior cross-population and cross-scenario adaptability, and practical feasibility for mobile-edge deployment.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Application Domains: Internet of Things, Sensor Networks & Smart Cities
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Web applications in cross-disciplinary domains and verticals such as mixed reality, smart cities, and digital health
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Stress affects physical and mental health, and wearable devices have been widely used to detect daily stress through physiological signals. However, these signals vary due to factors such as individual differences and health conditions, making generalizing machine learning models difficult. To address these challenges, we present Human Heterogeneity Invariant Stress Sensing (HHISS), a domain generalization approach designed to find consistent patterns in stress signals by removing person-specific differences. This helps the model perform more accurately across new people, environments, and stress types not seen during training. Its novelty lies in proposing a novel technique called person-wise sub-network pruning intersection to focus on shared features across individuals, alongside preventing overfitting by leveraging continuous labels while training. The study focuses especially on people with opioid use disorder (OUD)-a group where stress responses can change dramatically depending on their time of daily medication taking. Since stress often triggers cravings, a model that can adapt well to these changes could support better OUD rehabilitation and recovery. We tested HHISS on seven different stress datasets-four of which we collected ourselves and three public ones. Four are from lab setups, one from a controlled real-world setting, driving, and two are from real-world in-the-wild field datasets without any constraints. This is the first study to evaluate how well a stress detection model works across such a wide range of data. Results show HHISS consistently outperformed state-of-the-art baseline methods, proving both effective and practical for real-world use. Ablation studies, empirical justifications, and runtime evaluations confirm HHISS's feasibility and scalability for mobile stress sensing in sensitive real-world applications.

Problem

• Detect stress accurately across diverse individuals and conditions
• Generalize models despite signal variations from personal differences
• Adapt stress sensing for opioid users' rehabilitation needs

Innovation

• Domain generalization for stress signal consistency
• Person-wise sub-network pruning intersection technique
• Leveraging continuous labels to prevent overfitting