Arthur Loureiro Researcher
Contact
Name and title: Arthur LoureiroResearcher
ORCID0000-0002-4371-0876 Länk till annan webbplats.
Workplace: Department of Physics Länk till annan webbplats.
Visiting address Roslagstullsbacken 21, AlbaNova universitetscentrum
Postal address Fysikum106 91 Stockholm
About me
Dr. Arthur Loureiro is an observational cosmologist and astrostatistician at Stockholm University and the Oskar Klein Centre for Cosmoparticle Physics. He obtained his PhD in Astrophysics and Observational Cosmology from University College London, where he developed data-driven approaches to large-scale structure analyses. Following his doctorate, he held postdoctoral and research positions at University College London, Imperial College London, the University of Edinburgh, and Stockholm University, working at the interface of cosmological data analysis, statistical inference, and survey methodology.
His research focuses on multi-probe cosmological analyses that combine galaxy clustering, weak lensing, and cosmic microwave background observations within coherent Bayesian frameworks. He has contributed to constraints on neutrino properties from cosmology, including work establishing the first cosmological upper bound on the lightest neutrino mass by combining astrophysical and particle-physics data. More broadly, he develops harmonic-space analysis tools, hierarchical Bayesian models, and machine-learning–accelerated inference methods aimed at making large survey analyses robust, scalable, and reproducible.
Dr. Loureiro is actively involved in major international collaborations, including the Euclid Consortium, the LSST Dark Energy Science Collaboration (DESC), and the Kilo-Degree Survey (KiDS). He served as a Pipeline Scientist within LSST DESC and contributed to blind cosmological analysis efforts and cosmological likelihoods. He is also a Visiting Researcher at Imperial College London, where he maintains strong research collaborations through the Almanac Collaboration, focused on field-level inference and advanced Bayesian methods for cosmology. He currently leads a research programme supported by a Swedish National Space Agency Career Grant on synnergies between the Euclid Space Telescope and the Planck Satellite for measuring the lightest neutrino mass.
I am a lecturer on the following courses:
- Nuclear and Particle Physics, Astrophysics and Cosmology (FK5024)
- Statistical Methods in Physics (FK7061)
My research focuses on multi-probe cosmology, using observations of large-scale structure to test fundamental physics. By combining galaxy clustering, weak lensing, and cosmic microwave background data within coherent statistical frameworks, I develop methods to extract robust constraints on neutrino masses, dark energy, and extensions of the standard cosmological model.
A central theme of my work is neutrino cosmology. I study how massive neutrinos affect the growth of cosmic structure and how joint analyses of early- and late-Universe probes can break long-standing parameter degeneracies. My research contributed to establishing the first cosmological upper bound on the lightest neutrino mass through a consistent combination of cosmological observations and particle-physics constraints. This direction is currently supported by my Swedish National Space Agency Career Grant, which focuses on multi-survey neutrino inference combining Euclid, Planck, and particle physics experiments.
My methodology rests on three complementary pillars: multi-probe harmonic-space analyses, advanced Bayesian hierarchical modelling, and machine-learning acceleration for scalable inference.
Multi-Probe Cosmology and Blind Analysis
I integrate galaxy clustering, cosmic shear, galaxy–galaxy lensing, and CMB cross-correlations within unified inference pipelines. Within the LSST Dark Energy Science Collaboration (DESC), I serve as a Pipeline Scientist and lead blind cosmological analysis efforts, contributing to the design and validation of reproducible end-to-end workflows for Stage-IV survey data.
I am also actively involved in the Euclid Consortium and the Kilo-Degree Survey (KiDS), contributing to angular power spectrum methodologies, likelihood modelling, and joint multi-tracer analyses.
Field-Level Inference and Bayesian Hierarchical Modelling
A major strand of my research develops field-level inference frameworks that infer cosmological parameters directly from sky maps rather than compressed summary statistics. I am a member of the Aquila Consortium and contribute to large-scale Bayesian inference initiatives, including connections to BORG-style approaches.
I led the development of the Almanac framework and co-developed Flinch.jl, a differentiable extension that enables gradient-based inference from map-level data to cosmological parameters. These tools provide scalable hierarchical models suited to next-generation survey data.
Machine Learning and Large-Scale Structure Simulations
Machine learning plays a central role in accelerating computationally demanding analyses. I develop emulators for cosmological observables—particularly in models with massive neutrinos—allowing efficient exploration of high-dimensional parameter spaces.
In parallel, I co-develop the GLASS simulation library, a modular framework for forward-modelling galaxy surveys with realistic observational effects. GLASS supports validation of inference pipelines, multi-tracer analyses, and machine-learning–based acceleration strategies across several major cosmology collaborations.