Ipac logo
RINGS: A Calibrated Discovery Engine for Galaxy-Scale Strong Lenses in the First Two Years of Roman Public Imaging
Program ID 19011
Science Category Large Scale Structure of the Universe
Program Type Analysis
Category Medium
Principal Investigator Asantha Cooray
PI Institution University of California, Irvine
Co-Investigators
  • Cao Ye (University of California, Irvine)
  • Thomas Li (University of California, Irvine)
  • Xinmei Xu (University of California, Irvine)
  • Roger Chen (University of California, Irvine)
  • Ziqi Liu (University of California, Irvine)
Abstract Roman’s first two years of public Wide Field Instrument imaging will open a powerful new discovery space for galaxy-scale strong lenses. This program will build a calibrated, survey-scale lens-discovery engine for public Roman data, focused on doubles, quads, rings, and other galaxy-scale strong-lens configurations in HLWAS imaging and associated catalogs, with opportunistic use of eligible early time-domain products when available. The goal is not simply to rank candidates, but to deliver a statistically useful community resource: vetted candidate catalogs, benchmark cutout sets, uncertainty estimates, and calibrated selection-function products for downstream science. The scientific payoff is broad. A homogeneous Roman lens sample will enable studies of dark-matter substructure, lens demographics, rare high-magnification systems, and magnified high-redshift sources. The methodological payoff is equally important: Roman provides the ideal setting to move beyond first-generation CNN-only lens searches toward a hybrid discovery architecture that combines high-recall candidate generation, foundation-model reranking, structured expert vetting, and injection-recovery calibration. Public Euclid results already show that survey-scale lens mining is feasible, but also that human review becomes the dominant bottleneck at large scale. By concentrating true lenses into a smaller top-ranked subset, the proposed approach is designed to reduce expert-vetting cost while preserving scientific completeness and interpretability. The resulting products will turn early Roman public imaging into a calibrated strong-lens discovery resource for the community.