3 papers
stat.CO2026
A thermodynamic approach to Approximate Bayesian Computation with multiple summary statistics
Carlo Albert, Simone Ulzega, Simon Dirmeier +3
Bayesian inference with stochastic models is often difficult because their likelihood functions involve high-dimensional integrals. Approximate Bayesian Computation (ABC) avoids ev…
astro-ph.CO2026
Towards the Reconstruction of a Unified Dark Matter Halo: a Phenomenological Approach
Claudia Caputo, Daniele Bertacca, Alberto Bassi +1
We investigate static, spherically symmetric halo configurations within Unified Dark Matter (UDM) scalar-field models, developing a systematic mapping between standard cold dark ma…
cs.LG2026
When Bias Meets Trainability: Connecting Theories of Initialization
Alberto Bassi, Marco Baity-Jesi, Aurelien Lucchi +2
The statistical properties of deep neural networks (DNNs) at initialization play an important role to comprehend their trainability and the intrinsic architectural biases they poss…