The Shear-to-Cosmology Paradigm I. Hybrid Field-Level and Simulation-Based Framework for Weak Lensing Surveys

TL;DR

This paper introduces a hybrid machine-learning framework combining field-level inference and simulation-based inference to enhance cosmological parameter estimation from weak lensing shear data, outperforming traditional methods.

Contribution

The authors develop a novel hybrid ML approach that directly maps shear fields to cosmological parameters, incorporating PCA-based denoising and achieving significant improvements over standard techniques.

Findings

Shear-based inference doubles the Figure of Merit compared to convergence-based methods.

PCA denoising combined with ML compression improves FoM by 36.4%.

The framework is scalable and robust for Stage-IV weak-lensing surveys.

Abstract

Precise cosmological inference from next-generation weak lensing surveys requires extracting non-Gaussian information beyond standard two-point statistics. We present a hybrid machine-learning (ML) framework that integrates field-level inference (FLI) with simulation-based inference (SBI) to map observed shear fields directly to cosmological parameters, eliminating the need for convergence reconstruction. The FLI network extracts rich non-Gaussian information from the shear field to produce informative features, which are then used by SBI to model the resulting complex posteriors. To mitigate noise from intrinsic galaxy shapes, we develop a blind, training-free, PCA-based shear denoising method. Tests on CSST-like mock catalogs reveal significant performance gains. The shear-based inference achieves approximately twice the cosmological constraining power in Figure of Merit (FoM) compared to the conventional convergence-based approach. Moreover, the combination of PCA denoising and ML compression can deliver a 36.4% improvement in FoM over standard shear two-point statistics. This work establishes a scalable and robust pathway for cosmological inference, unlocking the full potential of Stage-IV weak-lensing surveys.