Recovery Maps and Approximate Markovianity
Small conditional mutual information guarantees that a tripartite state can be recovered approximately from its marginal by acting on . The theorem is quantitative and state dependent; it does not guarantee that the recovering channel is unique, geometrically local, causal, energy feasible, or easy to implement.
Required background. Use Conditional Mutual Information and Quantum Markov Structure. Helpful background. Markov generators and semigroups distinguish dynamical memory loss, while data processing supplies the information-loss perspective.
The recovery guarantee
Section titled “The recovery guarantee”For a finite-dimensional tripartite state, Fawzi and Renner proved that there exists a channel satisfying
where is the unsquared fidelity. Thus guarantees a recovered state with for natural logarithms. See Fawzi and Renner 2015, pp. 575–611.
The structural figure shows the one-way logic: correlation loss bounds attainable recovery, while choosing a concrete map is a further step.
Recovery translates small conditional mutual information into closeness of states. The bound asserts existence on a specified algebra; it does not supply spacetime locality, energy control, uniqueness, or an implementation cost. Schematic.
A regulated Gaussian chain
Section titled “A regulated Gaussian chain”Take three neighboring blocks in a finite harmonic chain. Compute from one covariance matrix and construct a Gaussian channel from to . Compare the recovered and exact covariance matrices and then their fidelity. This tests a restricted Gaussian recovery family; the theorem’s optimal channel may be non-Gaussian.
To claim a QFT limit, control the continuum embeddings and the energy carried by the recovered modes. A channel that reproduces ultraviolet correlations using unbounded squeezing can satisfy a formal fidelity target at each cutoff yet have no finite-energy limiting implementation.
Causal support is separate. An abstract operation on the algebra need not be generated inside the spacetime domain one intends to call local. The admissible operation algebra and time window must therefore be stated in an operational claim.
Existence versus feasibility
Section titled “Existence versus feasibility”The validity map separates the theorem’s hypotheses from common overinterpretations.
Changing the recovery domain, using postselection, ignoring rank support, or removing the energy bound changes the guarantee. Abstract recoverability is weaker than a local laboratory protocol. Schematic.
Report the recovered marginal, channel domain and codomain, fidelity convention, error, regulator, energy behavior, and any spacetime locality requirement. Without those data, “approximately Markov” is incomplete.
References
Section titled “References”- Fawzi, Omar, and Renato Renner. “Quantum Conditional Mutual Information and Approximate Markov Chains.” Communications in Mathematical Physics 340 (2015): 575–611. DOI. Open preprint.
Further reading
Section titled “Further reading”- Sutter, David, Omar Fawzi, and Renato Renner. “Universal Recovery Map for Approximate Markov Chains.” Proceedings of the Royal Society A 472 (2016): 20150623. DOI. Open preprint.