0 Assigned ID: 6128 ----- Attempting Submission 1 (Mon Apr 07 13:31:03 GMT 2025) ----- HST Phase I Proposal 6128 successfully submitted. Receipt: # 6128-1 ----- Attempting Submission 2 (Thu Apr 10 23:53:08 GMT 2025) ----- HST Phase I Proposal 6128 successfully submitted. Receipt: # 6128-2 ----- Attempting Submission 3 (Thu Apr 10 23:58:07 GMT 2025) ----- HST Phase I Proposal 6128 successfully submitted. Receipt: # 6128-3 ----- Attempting Submission 4 (Fri Apr 11 00:03:06 GMT 2025) ----- HST Phase I Proposal 6128 successfully submitted. Receipt: # 6128-4 ----- Attempting Submission 5 (Fri Apr 11 00:11:57 GMT 2025) ----- HST Phase I Proposal 6128 successfully submitted. Receipt: # 6128-5 ----- Attempting Submission 6 (Fri Apr 11 01:37:53 GMT 2025) ----- HST Phase I Proposal 6128 successfully submitted. Receipt: # 6128-6 ----- Attempting Submission 7 (Fri Apr 11 02:10:14 GMT 2025) ----- The nature of dark matter (DM) is critical to cosmology, yet cold dark matter (CDM) remains largely untested at halo masses below 10^13 M_Sun. This proposal outlines a novel approach to probe CDM predictions down to halo masses of ~10^11 M_Sun through archival analysis of high-resolution Hubble Space Telescope (HST) images. We aim to systematically discover gravitational lenses with small Einstein radii (0.03" <~ theta_E < 0.5"), using advanced machine learning methods to identify previously overlooked lenses. The proposed research specifically targets low halo mass (LHM) lenses, encompassing "deci-" and "centilensing," as a direct test of CDM predictions. Using a custom ResNet and a ResNet+U-Net pipeline, preliminary forecasts indicate we can detect a large number of lenses with nearly 100% purity, even in the unexplored centilensing regime. The discovery of these lenses will populate a critical gap--the "strong lensing desert"--in current catalogs of strong lenses. Additionally, unlike the discovery claims of "dark" (sub)halos, these systems can be unambiguously confirmed spectroscopically. Furthermore, we will perform lens modeling for confirmed systems, providing a rigorous statistical framework for testing CDM at unprecedentedly small scales (not counting the "dark" (sub)halo discovery claims). The results will significantly enhance future lensing studies with JWST and Roman, ultimately contributing insights into the fundamental properties of dark matter. PI Dr. Xiaosheng Huang is an expert in discovering strong lenses in imaging surveys and in strong lens modeling. He led several lens searches in ground-based imaging surveys using ML/AI techniques (e.g., Huang+ 2021, ApJ, 909, 27; Inchausti, Storfer, Huang+ 2025, in prep). He is the co-developer of the GIGA-Lens modeling code (Gu, Huang+ 2022 ApJ, 935, 49) and applied it to a strong lens with HST observations (Huang+2025, arxiv:2502.03455). Mr. Ethan Silver is an expert in using machine learning to find strong lenses of different sizes under different observational conditions with high purity (Silver, Wang, Huang, Bolton+ 2025, submitted). Mr. Richard Wang is an expert in applying segmentation to strong lensing systems and locating the smallest Einstein radius lenses with machine learning techniques. Dr. Adam Bolton has successfully conducted several previous spectroscopic lens surveys combining SDSS and HST observations. He is an expert in galaxy-scale strong gravitational lensing and in spectroscopic data analysis. Mr. Christopher Storfer is an expert in discovering strong lenses in imaging surveys and in DESI, Keck, and Gemini spectroscopy (Storfer+ 2025, in prep). Mr. Yuan-Ming Hsu has expertise in identifying low-mass lenses in imaging data and spectroscopy (Hsu+ 2025, in prep). Mr. Jose Inchausti has expertise in finding strong lenses using neural networks (Inchausti+ 2025, in prep). Mr. Harsh Ambardekar has expertise in lens modeling using GIGA-Lens for both ground- and space-based imaging. Dr. David Schlegel is co-Project Scientist of the DESI project. All Co-Is are members of the DESI collaboration, however this proposal only relies on DESI data that have already been publicly released. Galaxies 0 false false false false false true true SUPPORTED