Calculation of Auger spectra using the quantum generative eigensolver

Calculation of Auger spectra from first principles remains a difficult challenge on classical computers due to the need to resolve many excited states. We present here a computational pathway based on the newly-proposed generative quantum algorithm (GQE), in which a decoder-only transformer model inspired by GPT-2 is trained to generate quantum circuits for quantum chemical calculations. We believe this method is well-suited for large-sclae chemical calculations due to its use of GPU parallelization and the scalability of transformer models. We detail our computational workflow, and present results for a water molecule. Finally, we discuss generalization to systems larger than 40 qubits.