Calculating α -decay half-lives with artificial neural networks considering the effects of angular momentum and deformation

Artificial neural networks (ANNs) can be used to learn complex representations of data, enabling new approaches to modeling and processing in the physical sciences. In this work, ANNs are employed to calculate the α-decay half-lives of nuclei. An improvement in the predictive power of the ANN models can be achieved by incorporating the angular momentum transferred by α particle and the quadrupole deformation of parent nuclei. Consequently, the root-mean-square deviation between the ANN-predicted α-decay half-lives and the experimental data is reduced from 0.581 to 0.334. Predictions are made for the α-decay half-lives of isotopes with $Z=117$, 118, 119, and 120. Based on the characteristics (or systematics) of the α-decay half-lives, we propose that $N=184$ is a closed neutron shell beyond $N=126$.