{"id":"arxiv/2610.38028","url":"https://arxiv.org/abs/2610.38028","tags":["quant-ph","cond-mat.stat-mech"],"score":0,"title":"Sample Complexity for Generative Learning of Quantum Measurement Data","domain":"quant-ph","source":"arxiv","authors":["Shuhan Zhang","Yi-Zhuang You"],"reviews":0,"subject":"quant-ph","summary":"Generative learning of a quantum measurement distribution from finite data requires choosing a model capacity that balances representation accuracy against estimation cost. We develop a bias-variance framework that decomposes the expected population Kullback-Leibler error at model capacity $s$ and sample size $N$ as $\\mathcal E(s,N)=B(s)+V(s,N)$, with representation bias $B$ and a variance term $V$ capturing the mean estimation error. In the well-sampled regime, $V(s,N)\\simeq[NC(s)]^{-1}$, and exponential or algebraic laws for $B$ and $C$ yield four scaling templates for the optimal error,…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-10-27","reactions":{}}