Tackling LSST's sparse multiband light curves with Neural Processes

The sparse multiband light curves from LSST are crucial for informing follow-up observations and enabling the discovery of novel astronomical events. However, a challenge is the analysis of these light curves, with their long observational gaps and differing cadences over filters. While Gaussian Process (GP) regression is a common method for interpolation, it requires specifying a kernel a priori, making a strong assumption about the object's temporal nature that may not hold for diverse transients. However, the sheer volume of LSST light curves enables a data-driven approach. Neural Processes (NPs) are a family of models that learn distributions over functions, combining the probabilistic framework of GPs with the scalability and flexibility of neural networks. In this talk, I will discuss my ongoing work on the neural process family of models for light curve interpolation and its effect on downstream tasks like feature extraction, classification and anomaly detection.