Abstract
In this talk, a consistent database including nuclear mass, beta-decay rate, and beta-delayed neutron emission probability is introduced based on two deep neural network models. The prediction of nuclear mass from the neural network model is used to calculate the Q value and the neutron separation energy of beta decay, i.e., it is used as input of the models for beta decay process.
A deep neural network is developed for nuclear mass prediction, based on the finite-range droplet model (FRDM12). Different hyperparameters, including the activation function, the learning rate, the number of hidden units, and the initializers, are tested and adjusted in a systematic way. A deep neural network developed for the prediction of beta decay, based on the Gross theory. Both the beta decay rate and the beta delayed neutron emission probability are trained in the same neural network so that the number of data in the training set can be increased, and the relation between the beta decay rate and the beta delayed neutron emission probability can be shown in the model. With the neural network models used in this project, the accuracy of the prediction of nuclear mass, beta-decay rate, and beta-delayed neutron emission probability is improved.
Anyone interested is welcome to attend.