arXiv:2505.17354v3 Announce Type: replace Abstract: In many real-world settings--e.g., single-cell RNA sequencing, mobility sensing, and environmental monitoring--data are observed only as temporally aggregated snapshots collected over finite time windows, often with noisy or uncertain timestamps, and without access to continuous trajectories. We study the problem of estimating continuous-time dynamics from such snapshots. We present Continuous-Time Optimal Transport Flow (CT-OT Flow), a two-stage framework that (i) infers high-resolution time labels by aligning neighboring intervals via parti
This research addresses a fundamental challenge in data science: inferring continuous processes from discrete, noisy observations, a problem exacerbated by the increasing volume of temporal data in various fields.
Improved methods for extracting continuous dynamics from discrete data can significantly enhance predictive modeling, anomaly detection, and understanding of complex systems in fields like biology, sensing, and AI.
The CT-OT Flow framework proposes a more robust way to reconstruct continuous-time trajectories, potentially leading to more accurate simulations and deeper insights from sparse temporal data.
- · AI researchers
- · Bioinformatics
- · Environmental monitoring
- · Mobility sensing platforms
Researchers gain a more powerful tool for analyzing time-series data with inherent gaps and uncertainties.
Enhanced ability to model and predict complex dynamic systems across scientific and industrial applications.
This could accelerate discovery in fields relying on temporal data, leading to new insights and applications that were previously intractable.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG