Nonlocal growth of quantum conditional mutual information under decoherence

Yifan Zhang, Sarang GopalakrishnanarXiv 2024 · DOI · arXiv

TL;DR

This paper introduces quantum conditional mutual information (CMI) as a robust diagnostic for long-range correlations generated by local decoherence. The authors establish fundamental upper bounds on how fast CMI can grow and demonstrate that "teleportation" phase transitions—where local noise suddenly creates global correlations—belong to entirely different universality classes depending on whether the noise comes from measurements or erasures.

What Problem Does This Paper Address?

In the quantum world, local operations cannot create entanglement between distant regions. However, they can "unmask" or convert existing short-range entanglement into long-range entanglement. A classic example is Bell teleportation: by performing a local measurement on two qubits, you can entangle two other distant qubits.

Recently, researchers have been fascinated by "Measurement-Induced Entanglement" (MIE). The catch is that MIE is usually defined for a specific "trajectory"—a single set of measurement outcomes. If you average over all possible outcomes, the entanglement often vanishes. This paper seeks a more intrinsic, channel-based property to describe this phenomenon. They turn to Quantum Conditional Mutual Information (CMI). CMI measures the correlation between two regions ($A$ and $C$) given knowledge of a third ($B$). The authors ask: How much can CMI grow when the middle region $B$ undergoes decoherence, and what kind of quantum states are the "best" at facilitating this nonlocal growth?

Methods

The authors combine analytical mathematical proofs with numerical simulations:

  • Information-Theoretic Bounds: They derive rigorous upper bounds for CMI growth ($\delta I$) under general quantum channels and specific measurement channels using properties of Von Neumann entropy and the decoupling theorem.
  • State Characterization: They identify "efficient teleporters"—states that saturate these CMI growth bounds. They specifically examine Absolutely Maximally Entangled (AME) states and states related to the Hayden-Preskill black hole paradox.
  • Numerical Simulations: To study many-body dynamics, they use Random Clifford Circuits. These circuits are computationally efficient to simulate while capturing complex entanglement growth, allowing the authors to probe phase transitions in systems with up to 100+ qubits.
  • Statistical Mechanics Mapping: They map the quantum circuit dynamics to a classical Ising-like model to provide an intuitive "domain wall" explanation for the observed phase transitions.

Datasets

As a theoretical physics paper, it does not use external datasets. Instead, it generates data through:

  • Numerical simulation of 1D brickwork random Clifford circuits.
  • Mean-field approximations of the associated statistical mechanics models.

Key Insights and Ideas

  • The CMI Gain Bound: For a general channel, the CMI gain is bounded by twice the entropy added to the environment ($\delta I \leq 2S(E)$). For pure measurement channels, this bound is tighter: $\delta I \leq S(E)$.
  • Efficient Teleporters: Not all states are good at growing CMI. The authors show that "efficient teleporters" are states where different measurement outcomes result in post-selected states supported on orthogonal subspaces.
  • The Hayden-Preskill Connection: In states scrambled by Clifford unitaries (like the Hayden-Preskill model), measuring any subset of qubits in the middle region leads to maximal CMI growth. This makes the specific choice of which qubits to measure irrelevant—a property they call "efficient teleportation."
  • Erasure vs. Measurement: A major "aha moment" is that erasure channels (where qubits are lost) and measurement channels (where outcomes are known) behave differently. In 1D, the "teleportation transition" occurs at a circuit depth of $O(\log L)$ for measurements but requires a much deeper $L^{1/2}$ depth for erasures.

Key Results

  • Phase Transitions: The paper identifies a critical circuit depth above which CMI growth saturates its theoretical maximum. This is interpreted as a "teleportation phase transition."
  • Universality Classes: They argue that measurements and erasures belong to different universality classes. Specifically, the lower critical dimension for the teleportation transition is higher for erasures ($d=3$) than for measurements ($d=2$).
  • Quantitative Scaling: In 1D, they numerically confirm that the correlation length for teleported information grows as $\xi(t) \sim t$ for measurements but follows a different scaling for erasures due to the "random-field" nature of the mapping.

Limitations and Future Work

  • Qubit Focus: While the information-theoretic results are general, the numerical and circuit-based results focus primarily on qubit systems.
  • Clifford vs. Haar: Much of the efficient simulation relies on Clifford circuits. While these are representative, generic "Haar-random" unitaries (which are more physically "scrambling") show sub-maximal CMI growth, as explored in the appendices.
  • Future Directions: The authors suggest applying CMI growth as a diagnostic for more complex phases of matter, such as topological phases or cluster states, and investigating how these transitions manifest in experimental platforms like trapped ions or superconducting qubits where "mid-circuit" measurements are now possible.

Summary updated

Abstract

Local measurements cannot create entanglement, but they can convert short-range entanglement to long-range entanglement, as in quantum teleportation. This phenomenon of measurement-induced entanglement (MIE) has been widely discussed in recent work on measurement-induced entanglement phase transitions and related phenomena. Here, we situate MIE in a broader context of the growth of long-range conditional mutual information (CMI) under decoherence. We upper-bound the rate at which decoherence can generate long-range CMI, and derive a characterization of states that saturate this bound. We point out that the structure of states saturating the CMI upper bound can be very different under different decoherent dynamics and provide explicit examples. We additionally explore the dynamics of CMI in random quantum circuits subject to random local decoherence, as a function of circuit depth. We argue that the universality class of the finite-depth teleportation transition, as well as its lower critical dimension, are different for erasures than for measurements.
BibTeX
@misc{zhang2024nonlocal,
  author = {Yifan Zhang and Sarang Gopalakrishnan},
  title = {{Nonlocal growth of quantum conditional mutual information under decoherence}},
  year = {2024},
  eprint = {2402.03439},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2402.03439},
}
Figure