MEASURE: Multi-scale Minimal Sufficient Representation Learning for Domain Generalization in Sleep Staging

2025-10-13

AI Summary

Deep learning models exhibit poor generalization in cross-subject sleep staging, primarily due to physiological signal distribution shifts and the inability of existing domain generalization methods—such as contrastive learning—to adequately disentangle domain-specific redundant information, leading to overfitting in high-level features and underutilization of multi-scale time-frequency patterns. To address this, we propose a Multi-Scale Minimal Sufficient Representation Learning framework that explicitly disentangles and suppresses domain-related redundancy across multiple feature hierarchies while preserving discriminative temporal and spectral structures. By integrating multi-scale contrastive learning with an information bottleneck constraint, our method achieves robust domain-invariant representation learning. Extensive cross-domain experiments on SleepEDF-20 and MASS demonstrate that our approach significantly outperforms current state-of-the-art methods, achieving average Kappa improvements of 3.2–5.7 percentage points.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Representation Learning for Vision
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs

Abstract

Deep learning-based automatic sleep staging has significantly advanced in performance and plays a crucial role in the diagnosis of sleep disorders. However, those models often struggle to generalize on unseen subjects due to variability in physiological signals, resulting in degraded performance in out-of-distribution scenarios. To address this issue, domain generalization approaches have recently been studied to ensure generalized performance on unseen domains during training. Among those techniques, contrastive learning has proven its validity in learning domain-invariant features by aligning samples of the same class across different domains. Despite its potential, many existing methods are insufficient to extract adequately domain-invariant representations, as they do not explicitly address domain characteristics embedded within the unshared information across samples. In this paper, we posit that mitigating such domain-relevant attributes-referred to as excess domain-relevant information-is key to bridging the domain gap. However, the direct strategy to mitigate the domain-relevant attributes often overfits features at the high-level information, limiting their ability to leverage the diverse temporal and spectral information encoded in the multiple feature levels. To address these limitations, we propose a novel MEASURE (Multi-scalE minimAl SUfficient Representation lEarning) framework, which effectively reduces domain-relevant information while preserving essential temporal and spectral features for sleep stage classification. In our exhaustive experiments on publicly available sleep staging benchmark datasets, SleepEDF-20 and MASS, our proposed method consistently outperformed state-of-the-art methods. Our code is available at : https://github.com/ku-milab/Measure

Problem

• Addressing domain generalization challenges in sleep staging models
• Reducing domain-relevant information while preserving essential features
• Improving performance on unseen subjects through multi-scale representation learning

Innovation

• Multi-scale minimal sufficient representation learning framework
• Reduces domain-relevant information while preserving features
• Leverages diverse temporal and spectral information across levels