Description
The forthcoming high-quality data from the next generation of galaxy surveys promises powerful new tests of cosmology. However, fully realizing this potential poses new challenges in theoretical modeling and analysis. Novel analysis techniques are required to maximize the cosmological information extracted from the data while remaining robust to theoretical uncertainties in the small-scale processes of galaxy formation. This talk introduces perturbative forward-modeling of the large-scale galaxy density with LEFTfield. Such a differentiable forward model enables the analysis of galaxy clustering at the level of density fields. Forgoing the compression of the data into summary statistics, field-level analysis can maximize the extracted cosmological information and provide insights into the initial conditions and the structure formation history over the survey volume. The perturbative approach, particularly the description of galaxy bias with the Effective Field Theory of Large-Scale Structure (EFTofLSS), allows for robust marginalization over theoretical uncertainties. Moreover, forward-modeling the process of structure allows for the self-consistent incorporation of many observational and systematic effects, with redshift-space distortions being one concrete example. Finally, the talk will outline forthcoming challenges and highlight some recent applications of LEFTfield.
