Speaker
Description
The search for viable alternatives to Lambda-CDM requires not only new gravitational models, but also systematic criteria to identify which regions of their parameter space are physically meaningful and observationally testable. In this talk, I present a cosmological model based on non-metricity gravity (f(Q, Φ)), coupled to a non-canonical dilatonic scalar field, and discuss how dynamical systems can be used as a theoretical guide for Bayesian parameter inference. Starting from the effective action, the modified Friedmann equations are derived and rewritten as an autonomous system in terms of suitable dimensionless variables. The critical-point structure is then analyzed in order to identify radiation-dominated, matter-dominated, scaling, and late-time accelerated solutions. The existence conditions, stability properties, eigenvalues, and effective cosmological parameters associated with these critical points provide a first viability filter for the model. In particular, they indicate which parameters directly control the cosmological dynamics and are therefore expected to be constrained by observational data. A central point of the methodology is the construction of physically motivated priors for Markov Chain Monte Carlo analyses. Instead of choosing arbitrary numerical intervals, the prior ranges are selected through three complementary criteria. First, one identifies the model parameters that explicitly appear in the coordinates, existence conditions, eigenvalues, or effective observables of the relevant dark-energy critical points. Parameters that do not enter this dynamical structure are expected to generate weakly informative or degenerate marginal distributions and may be fixed or treated as secondary hyperparameters. Second, the prior intervals are refined by numerically evolving the autonomous system and requiring convergence toward the desired cosmological attractor, as seen in the evolution of the dimensionless variables with the number of N e-folds. Third, the resulting viable regions are tested at the background level by comparing the predicted Hubble function (H(z)) with the Lambda-CDM curve, requiring the prior extremes to explore controlled deviations around the standard cosmological expansion. Once these dynamically informed priors are established, the model is confronted with observational expansion data, including Cosmic Chronometers, Type Ia supernovae, and BAO/DESI measurements. The resulting MCMC contours are then interpreted not only as statistical constraints, but also as observational projections of the underlying phase-space structure. In this sense, correlations and degeneracies in the corner plots can be related to the existence and stability of the critical points that govern the cosmic evolution. This work aims to provide a transparent connection between modified gravity, dynamical stability, prior selection, and observational inference. The proposed framework shows how critical points can be used not only to classify the qualitative evolution of a cosmological model, but also to guide the construction of priors and to interpret which parameters are genuinely constrained by cosmological data.