1 December 2021 in Computer Physics Communications
1 December 2021 in Physics Letters B
13 July 2022 in Physical Review C
Machine learning methods and uncertainty quantification have been gaining interest throughout the last several years in low-energy nuclear physics. In particular, Gaussian processes and Bayesian neural networks have increasingly been applied to improve mass model predictions while providing well-quantified uncertainties. In this work, we use the probabilistic Mixture Density Network (MDN) to directly predict the mass excess of the 2016 Atomic Mass Evaluation within the range of measured data, and we extrapolate the inferred models beyond available experimental data. The MDN provides not only mean values but also full posterior distributions both within the training set and extrapolated testing set. We show that the addition of physical information to the feature space increases the accuracy of the match to the training data as well as provides for more physically meaningful extrapolations beyond the the limits of experimental data.
1 August 2022 in Physical Review C
We present an approach to modeling the ground-state mass of atomic nuclei based directly on a probabilistic neural network constrained by relevant physics. Our physically interpretable machine learning (PIML) approach incorporates knowledge of physics by using a physically motivated feature space in addition to a soft physics constraint that is implemented as a penalty to the loss function. We train our PIML model on a random set of approximately 20% of the atomic mass evaluation (AME) and predict the remaining 80%. The success of our methodology is exhibited by a keV match to data for the training set and keV for the entire AME with . We show that our general methodology can be interpreted using feature importance.
24 November 2021 in Physical Review Letters
We investigate the angular momentum removal from fission fragments (FFs) through neutron and -ray emission, finding that about half the neutrons are emitted with angular momenta and that the change in angular momentum after the emission of neutrons and statistical rays is significant, contradicting usual assumptions. Per fission event, in our simulations, the neutron and statistical -ray emissions change the spin of the fragment by , with a large standard deviation comparable to the average value. Such wide angular momentum removal distributions can hide any underlying correlations in the fission fragment initial spin values. Within our model, we reproduce data on spin measurements from discrete transitions after neutron emissions, especially in the case of light FFs. The agreement further improves for the heavy fragments if one removes from the analysis the events that would produce isomeric states. Finally, we show that while in our model the initial FF spins do not follow a sawtoothlike behavior observed in recent measurements, the average FF spin computed after neutron and statistical emissions exhibits a shape that resembles a sawtooth. This suggests that the average FF spin measured after statistical emissions is not necessarily connected with the scission mechanism as previously implied.
27 January 2021 in Physical Review C
The Hauser-Feshbach fission fragment decay model, , which calculates the statistical decay of fission fragments, has been expanded to include multichance fission, up to neutron incident energies of 20 MeV. The deterministic decay takes as input prescission quantities—fission probabilities and the average energy causing fission—and postscission quantities—yields in mass, charge, total kinetic energy, spin, and parity. From these fission fragment initial conditions, the full decay is followed through both prompt and delayed particle emissions, allowing for the calculation of prompt neutron and properties, such as multiplicity and energy distributions, both independent and cumulative fission yields, and delayed neutron observables. In this work, we describe the implementation of multichance fission in the model, and show an example of prompt and delayed quantities beyond first-chance fission, using the example of neutron-induced fission on . This expansion represents significant progress in consistently modeling the emission of prompt and delayed particles from fissile systems.