Immediate plans

This repo is home to a series of LSST-DESC projects aiming to quantify the impact of imperfect prior information on probabilistic redshift estimation. An outline of the baseline RAIL is illustrated here.

1. Golden Spike: Build the basic infrastructure for controlled experiments of forward-modeled photo-z posteriors

  • a rail.creation subpackage that can generate true photo-z posteriors and mock photometry.

  • a rail.estimation subpackage with a superclass for photo-z posterior estimation routines and at least one subclass template example implementing the trainZ (experimental control) algorithm.

  • a rail.evaluation subpackage that calculates at least the metrics from the PZ DC1 Paper for estimated photo-z posteriors relative to the true photo-z posteriors.

  • documented scripts that demonstrate the use of RAIL in a DC1-like experiment on NERSC.

  • sufficient documentation for a v1.0 release.

  • an LSST-DESC Note presenting the RAIL infrastructure.

2. RAILroad: Quantify the impact of nonrepresentativity (imbalance and incompleteness) of a training set on estimated photo-z posteriors by multiple machine learning methods

  • parameter specifications for degrading an existing Creator to make an imperfect prior of the form of nonrepresentativity into the observed photometry.

  • at least two Estimator wrapped machine learning-based codes for estimating photo-z posteriors.

  • additional Evaluator metrics with feed-through access to the qp metrics.

  • end-to-end documented scripts that demonstrate a blinded experiment on NERSC.

  • an LSST-DESC paper presenting the results of the experiment.