Omni-Sleep: A Sleep Foundation Model via Hierarchical Contrastive Learning of CNS--ANS Dynamic
Updated
Updated · arxiv.org · Jul 9
Omni-Sleep: A Sleep Foundation Model via Hierarchical Contrastive Learning of CNS--ANS Dynamic
1 articles · Updated · arxiv.org · Jul 9
Summary
Researchers have introduced Omni-Sleep, a foundation model that leverages CNS and ANS dynamics to improve sleep analysis using multimodal polysomnography data.
Omni-Sleep employs hierarchical contrastive learning to capture both intra-system consistency and inter-system synchronization, resulting in robust, generalizable sleep representations.
Pre-trained on over 100,000 hours of data, Omni-Sleep outperforms existing models in sleep staging and disease classification, especially under label scarcity and missing modalities.