A Poisson Process AutoDecoder for X-ray Sources
This work addresses the problem of analyzing X-ray astronomical data for researchers, offering a novel method that directly captures Poisson processes, though it is incremental in applying neural fields to this domain.
The paper tackles the challenge of modeling X-ray source data, which follows a Poisson process with high variability, by introducing the Poisson Process AutoDecoder (PPAD) to reconstruct rate functions and generate representations, achieving improvements in tasks like classification and anomaly detection on the Chandra Source Catalog.
X-ray observing facilities, such as the Chandra X-ray Observatory and the eROSITA, have detected millions of astronomical sources associated with high-energy phenomena. The arrival of photons as a function of time follows a Poisson process and can vary by orders-of-magnitude, presenting obstacles for common tasks such as source classification, physical property derivation, and anomaly detection. Previous work has either failed to directly capture the Poisson nature of the data or only focuses on Poisson rate function reconstruction. In this work, we present Poisson Process AutoDecoder (PPAD). PPAD is a neural field decoder that maps fixed-length latent features to continuous Poisson rate functions across energy band and time via unsupervised learning. PPAD reconstructs the rate function and yields a representation at the same time. We demonstrate the efficacy of PPAD via reconstruction, regression, classification and anomaly detection experiments using the Chandra Source Catalog.