English

Deconstruction and conditional erasure of quantum correlations

Quantum Physics 2018-10-24 v2 Information Theory math.IT

Abstract

We define the deconstruction cost of a tripartite quantum state on systems ABEABE as the minimum rate of noise needed to apply to the AEAE systems, such that there is negligible disturbance to the marginal state on the BEBE systems, while the system AA of the resulting state is locally recoverable from the EE system alone. We refer to such actions as deconstruction operations and protocols implementing them as state deconstruction protocols. State deconstruction generalizes Landauer erasure of a single-party quantum state as well the erasure of correlations of a two-party quantum state. We find that the deconstruction cost of a tripartite quantum state on systems ABEABE is equal to its conditional quantum mutual information (CQMI) I(A;BE)I(A;B|E), thus giving the CQMI an operational interpretation in terms of a state deconstruction protocol. We also define a related task called conditional erasure, in which the goal is to apply noise to systems AEAE in order to decouple system AA from systems BEBE, while causing negligible disturbance to the marginal state of systems BEBE. We find that the optimal rate of noise for conditional erasure is also equal to the CQMI I(A;BE)I(A;B|E). State deconstruction and conditional erasure lead to operational interpretations of the quantum discord and squashed entanglement, which are quantum correlation measures based on the CQMI. We find that the quantum discord is equal to the cost of simulating einselection, the process by which a quantum system interacts with an environment, resulting in selective loss of information in the system. The squashed entanglement is equal to half the minimum rate of noise needed for deconstruction/conditional erasure if Alice has available the best possible system EE to help in the deconstruction/conditional erasure task.

Cite

@article{arxiv.1609.06994,
  title  = {Deconstruction and conditional erasure of quantum correlations},
  author = {Mario Berta and Fernando G. S. L. Brandao and Christian Majenz and Mark M. Wilde},
  journal= {arXiv preprint arXiv:1609.06994},
  year   = {2018}
}

Comments

21 pages, 3 figures, accepted for publication in Physical Review A

R2 v1 2026-06-22T15:57:59.837Z