{"id":"arxiv/2610.17755","url":"https://arxiv.org/abs/2610.17755","tags":["quant-ph","cs.ET","cs.MS"],"score":0,"title":"PADO-Pauli: A GPU-Accelerated Implementation of Pauli Propagation","domain":"quant-ph","source":"arxiv","authors":["Hyunwoo Kim","Youngseok Lee"],"reviews":0,"subject":"quant-ph","summary":"Variational quantum algorithms utilize parameterized circuits many times during training, at new parameter values, over batches of data, and with gradients. Pauli propagation approximates the expectation values of such circuits classically by evolving the observable in the Heisenberg picture as a truncated sum of Pauli strings. We present PADO-Pauli, a GPU engine for Pauli propagation that compiles a circuit and its observables once into sparse matrices in which each rotation angle contributes only a cosine and a sine factor, so every later evaluation reuses the compiled program. It…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-10-13","reactions":{}}