Legacy Survey of Space and Time cadence strategy evaluations for active galactic nucleus time-series data in Wide-Fast-Deep field

ABSTRACT Machine learning is a promising tool to reconstruct time-series phenomena, such as variability of active galactic nuclei (AGNs), from sparsely sampled data. Here, we use three Continuous Autoregressive Moving Average (CARMA) representations AGN – the Damped Random Walk (DRW) and (over/under)Damped Harmonic Oscillator simulate 10-yr light curves they would appear in upcoming Vera Rubin Observatory Legacy Survey Space Time (LSST), provide public generate these for any survey cadence. We investigate impact on science five proposed cadence strategies LSST’s primary Wide-Fast-Deep (WFD) survey. apply first time astronomy novel Stochastic Recurrent Neural Network (SRNN) algorithm input simulated LSST data, metric evaluate how well SRNN can help recover underlying CARMA parameters. find that light-curve reconstruction most sensitive duration gaps between observing season, cadences, those change balance filters, or avoid having long g band perform better. Overall, means densely long-term structure function DRW process (SF∞) reasonably well. However, all CARMA/SRNN models struggle decorrelation time-scale (τ) due observations. This may indicate major limitation using WFD data science.