{"id":"arxiv/2609.27128","url":"https://arxiv.org/abs/2609.27128","tags":["quant-ph"],"score":0,"title":"Efficient Learning of Clifford-Scrambled Product States","domain":"quant-ph","source":"arxiv","authors":["Tobias Haug"],"reviews":0,"subject":"quant-ph","summary":"Clifford circuits acting on product magic states provide a compact ansatz exhibiting extensive magic, volume-law entanglement, and even classically hard sampling under standard complexity assumptions. Here, we efficiently recover its hidden product subsystems from two-copy Bell sampling. Quadratic relations between Bell samples determine the irreducible blocks after removing Pauli stabilizers, and binary linear algebra constructs a Clifford disentangler. For logarithmic-size blocks, approximate learning of the full state is efficient whenever the state remains inverse-polynomially separated…","category":"research","comments":0,"disputed":false,"division":"physics","mentions":0,"discussed":0,"published":"2026-09-22","reactions":{}}