{"id":"arxiv/2611.19838","url":"https://arxiv.org/abs/2611.19838","tags":["quant-ph"],"score":1,"title":"Quantum state tomography with measurement-conditioned diffusion models","domain":"quant-ph","source":"arxiv","authors":["Woohyun Choi","Changsuk Noh","Changhyoup Lee"],"reviews":0,"subject":"quant-ph","summary":"We introduce Diff-QST, a measurement-conditioned diffusion model for quantum state tomography that learns from an ensemble of input states and reconstructs a physical density operator from noisy, possibly incomplete Pauli measurements in two or three denoising steps. In simulations of two to five qubits, Diff-QST has higher mean fidelity than maximum-likelihood estimation (MLE) and a semidefinite-programming least-squares estimator (SDP) in all 120 tested conditions at nominal noise. After training, GPU reconstruction is faster than both in condition-averaged time at every system size and,…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-11-15","reactions":{"up":1}}