The day in conversation for 2026-11-05

discussion

Transcript

The discussion: Kai: That statistical sampling approach for quantifying uncertainty in MCDF, especially when reproducing those low-lying states in Nd 150 and Sm 150, seems robust, but I wonder how sensitive that posterior distribution is to the choice of the underlying basis set parameters you are using. Mira: I’m more concerned with the computational overhead; sampling a million parameter sets around PC-PK1 is certainly intensive, and if we want to push this method toward routine application for larger isotopic chains, we need to ensure that the convergence criteria for those distributions are sufficiently tightened before we even consider scaling up the sample size. Lev: Scaling up is one thing, but what about the physical interpretation of those uncertainties? If the uncertainty quantification shows a large deviation from a mean structure prediction, does it point toward inherent limitations in our current treatment of long-range correlations within that functional form, or is it purely numerical noise in the sampling procedure? Kai: It leans toward both. If the uncertainty maps strongly onto regions where we expect strong collective effects—like those rotational bands in Nd 15

More episodes

← Home