Speaker
Description
Recent cosmological observations suggest possible deviations from a cosmological constant, pointing toward a dynamical nature of dark energy. Quintessence models, which assume a slowly rolling scalar field, provide a compelling theoretical framework to explain this late time evolution in the dark energy equation of state. However, identifying the correct form of the quintessence potential remains a major challenge, due to both theoretical constraints and the vast landscape of functional possibilities. In my work, I explore the use of symbolic regression, an interpretable machine learning technique, to discover viable quintessence potentials directly from observational data. By searching over analytical expressions rather than fitting predefined forms, symbolic regression offers a data-driven approach to model selection that retains physical interpretability.