Detecting the chiral magnetic effect via deep learning
Yuan-Sheng Zhao, Lingxiao Wang, Kai Zhou, Xu-Guang Huang
Abstract
The search for chiral magnetic effect (CME) in heavy-ion collisions has attracted long-term attention. Multiple observables are proposed, but all suffer large background contaminations. In this study, we construct an end-to-end CME-meter based on a deep convolutional neural network. After being trained over a dataset generated by a multiphase transport model, the CME-meter shows high accuracy in recognizing the CME-featured charge separation from the final-state pion spectra. It also exhibits remarkable robustness to diverse conditions including different collision energies, centralities, and elliptic flow backgrounds. In extrapolation tests, the CME-meter is validated in isobaric collisions and different model simulation, showing robust generalization performance. Moreover, based on variational approaches, we utilize the DeepDream method to derive the most responsive CME-spectrum that demonstrates the physical contents the machine learned.