Type: Poster
Session: Posters (Monday & Tuesday)
Author: Jonathan Calixto

Abstract: Chromatic biases in the point spread functions (PSFs) will bias measurements of cosmic shear at an unacceptable level in Rubin data if not corrected or calibrated. We explore an unsupervised machine-learning technique (self-organizing map, or SOM) that can be used to reduce a high dimensional data set to lower dimensions. The SOM facilitates visualization and exploration of correlations with features that may or may not have been used in the training. In our work, the high dimensional space is all the colors measured in the LSST bands, and the correlated feature is the chromatic bias in the PSF.  We measure how well the cell-position in a trained SOM can predict the chromatic bias in the PSF size or shape.