The PHAST survey is creating a legacy map of the southern half of M31’s star-forming disk (Chen et al., in prep) with photometric data from Hubble Space Telescope (HST) observations in four filters, using the Advanced Camera for Surveys (ACS) in the optical (F475W, F814W) and the Wide Field Camera 3 (WFC3) in the near ultraviolet (F275W, F336W). As the closest large metal-rich spiral to the Milky Way, M31 offers a unique laboratory to resolve stars in the disk to study star and galaxy formation and evolution. The PHAST survey, spanning ~13 kpc along the southern major axis, covers regions structurally unique and more indicative of merger history than the northern disk, as mapped in the completed PHAT program. Over 98% of the entire PHAST observations are now complete. We have measured stellar photometry for roughly 100 million resolved stars in M31, using simultaneous point-spread function (PSF) fitting across all bands and source positions, following advanced cosmic rays removal and precise astrometric image alignment. These observations will enable us to refine our understanding of the merger and accretion history of M31, as well as the formation and evolution of its disk.

deepCR is a deep-learning-based cosmic ray (CR) rejection framework originally presented by Zhang & Bloom (2020). The original approach requires a dedicated training set that consists of multiple frames of the same fields, enabling automatic CR labeling through comparison with their median co-adds. Here, we present a novel training approach that circumvents the need for a dedicated training set, but instead utilizes dark frames and the science images requiring CR removal themselves. During training, CRs present in dark frames are added to the science images, which the network is then trained to identify. In turn, the trained deepCR model can then be applied to identify CRs originally present in the science images. Using this approach, we present a new deepCR model trained on a diverse set of Hubble Space Telescope (HST) images taken from resolved galaxies in the local group, which is universally applicable across all WFC3/UVIS filters. We further evaluate the performance of the deepCR model on the Panchromatic Hubble Andromeda Southern Treasury (PHAST) survey and provide insights into how users can robustly determine the threshold for generating binary cosmic-ray masks from deepCR probability map predictions. This work has been submitted to ApJS.