aiqbench analysis
This research advances medical signal processing by integrating cyclostationarity directly into diffusion models, explicitly encoding periodic structures in cardiac and brain recordings to improve restoration under noise. By combining shift-covariant wavelet representations with phase conditioning, the method reduces sampling variance and network evaluations while maintaining diagnostic accuracy. The work demonstrates that leveraging measurable cyclic properties offers a significant efficiency gain over implicit learning approaches in physiological time-series analysis.
Cyclostationary Phase Conditioning for Medical Time Series Diffusion
arXiv:2609.34965v1 Announce Type: new Abstract: Many physiological time series, such as cardiac and brain recordings, exhibit cyclostationarity: their statistics vary periodically with an underlying cycle phase. Corruption from motion, poor contact, and physiological interference obscures morphology needed for diagnosis, making signal restoration essential. Existing diffusion approaches condition on corrupted observations alone and must learn cyclic structure implicitly. We instead propose two inductive biases which encode cyclostationarity: a shift-covariant wavelet representation and dense per-sample phase conditioning inferred from the corrupted input. We further introduce a training-free cyclostationarity index that quantifies phase structure and predicts when phase conditioning will help. Finally, we propose antithetic coupling of reverse trajectories to reduce sampling variance while achieving comparable performance with fivefold fewer network evaluations. Across modalities, our results show that explicitly encoding measurable cyclic structure improves physiological time-series restoration.
Abstract:Many physiological time series, such as cardiac and brain recordings, exhibit cyclostationarity: their statistics vary periodically with an underlying cycle phase. Corruption from motion, poor contact, and physiological interference obscures morphology needed for diagnosis, making signal restoration essential. Existing diffusion approaches condition on corrupted observations alone and must learn cyclic structure implicitly. We instead propose two inductive biases which encode cyclostationarity: a shift-covariant wavelet representation and dense per-sample phase conditioning inferred from the corrupted input. We further introduce a training-free cyclostationarity index that quantifies phase structure and predicts when phase conditioning will help. Finally, we propose antithetic coupling of reverse trajectories to reduce sampling variance while achieving comparable performance with fivefold fewer network evaluations. Across modalities, our results show that explicitly encoding measurable cyclic structure improves physiological time-series restoration.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Signal Processing (eess.SP)
I.2.6; I.5.4; J.3; G.3
arXiv:2609.34965 [cs.LG]
(or arXiv:2609.34965v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2609.34965 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 233 words
