This session will start with a short introduction to the Rubin Data Management team’s PZ Roadmap, including an update from the PZ Commissining Team, and short status updates on PZ-related software infrastructure. Then the focus will be on recent advances in photo-z algorithms, estimators, and performance. All are welcome. To contribute a talk in this session, please contact the session chair.

Preliminary Agenda

• 2 min -- Melissa Grahm -- Welcome
• 10 min -- Jeffrey Newman -- PZ for Next-Generation Surveys
• 8 min -- Melissa Grahm -- Update on the PZ Roadmap & Commissioning
• 10 min -- Sam Schmidt -- Update from the DESC PZ Working Group
• 10 min -- Simona Mei -- Getting image-based ML PZ estimators working with RAIL
• 10 min -- Julia Gschwend -- Update from BRA-LIN S4 PZ Services for LSST
• 10 min -- Renée Hlozek -- Photometric redshifts and supernova cosmology
• 10 min -- Biprateep Dey -- Calibrated Predictive Distributions for Photometric Redshifts (abstract below)
• 10 min -- Alex Malz -- Discussion: commissioning/XSC pipeline resources (open Q&A)

Abstracts

Update from BRA-LIN S4 PZ Services for LSST (Julia Gschwend)
As part of the LSST in-kind contribution program, LIneA will host an Independent Data Access Center (Lite-IDAC) and provide software infrastructure as a service to support the production of Photometric Redshifts on the LSST scale. In this talk, I will briefly describe the two main services, PZ Server and PZ Compute, and give an update on the development work done since the last presentation at the PZ Virtual Symposium. Please find more details on the BRA-LIN-S4 description page.

Photometric redshifts and supernova cosmology (Renée Hlozek)
Photometric supernova cosmology will be one of the cornerstones of dark energy science with Rubin. Fitting the magnitudes to light curve templates provides a distance modulus and associated cosmology constraints (depending on the type of supernova observed), however a crucial ingredient is the redshift of the supernova host.  I'll present some of the challenges of using photometric host redshifts on supernova cosmology, including tests from the Photometric LSST Astronomical Time-Series Classification Challenge (PLAsTiCC), and highlight a path forward.

Calibrated Predictive Distributions for Photometric Redshifts (Biprateep Dey, University of Pittsburgh, USA)
Many astrophysical analyses depend on estimates of redshifts (a proxy for distance) determined from photometric (i.e., imaging) data alone. Inaccurate estimates of photometric redshift uncertainties can result in large systematic errors. However, probability distribution outputs from many photometric redshift methods do not follow the frequentist definition of a Probability Density Function (PDF) for redshift — i.e., the fraction of times the true redshift falls between two limits z1 and z2 should be equal to the integral of the PDF between these limits. Previous works have used the global distribution of Probability Integral Transform (PIT) values to re-calibrate PDFs, but offsetting inaccuracies in different regions of feature space can conspire to limit the efficacy of the method. We leverage a recently developed regression technique that characterizes the local PIT distribution at any location in feature space to perform a local re-calibration of photometric redshift PDFs resulting in calibrated predictive distributions. Though we focus on an example from astrophysics, our method can produce predictive distributions which are calibrated at all locations in feature space for any use case.