Abstract:
   Weak lensing convergence maps provide a direct view of the projected matter distribution and underpin a broad range of cosmological analyses, from two-point power spectra to non-Gaussian statistics and multi-probe cross-correlations. Reconstructing these maps from masked, noisy shear catalogs remains a challenging inverse problem: standard Kaiser-Squires inversion is biased under realistic survey conditions, while prior-dependent methods (Wiener filtering, sparsity priors, diffusion models) alter the statistical properties of the recovered field. We present the Accurate Kappa Reconstruction Algorithm (AKRA) program, which provides the prior-free, maximum-likelihood solution to the mass-mapping problem on the curved sky. We report the first real-data application of AKRA to HSC Y1, producing unbiased convergence maps across all six HSC Y1 fields, and present the application of AKRA 3.0 to the DES Y3 Metacalibration shear catalog, yielding the highest-resolution, prior-free convergence map of the DES Y3 footprint to date.