Skip to content

Non-stationary Periodic LDP Reconstruction

Home / Open Questions / Non-stationary Periodic LDP Reconstruction

Background: Periodic signals collected under local differential privacy (LDP) often exhibit non-stationary behaviors such as tempo drift, which complicates long-term synchronization and reconstruction.

Question / Future Work: Current reconstruction frameworks for periodic time series under LDP primarily focus on stationary signals, making it difficult to maintain reconstruction quality when the underlying period length or phase alignment is dynamic. Developing robust techniques for non-stationary signals remains an open research challenge.

Why It Matters: Addressing non-stationary periodicity is critical for applying these privacy-preserving techniques to real-world multimedia streams, such as audio or heart rate, where temporal dynamics are intrinsic.

Evidence: These gaps motivate a period-conscious view of LDP reconstruction… [the framework is designed for] heterogeneous periodic structures common in real-world multimedia, such as variable period lengths, tempo drift, and multi-period superposition.

Metadata & Links