Entanglement dynamics and Page curves in random permutation circuits

This paper investigates entanglement dynamics in random permutation circuits, deriving tight upper bounds on entanglement generation based on initial state properties and demonstrating that while finite-size entanglement differs from that of random unitary circuits, their Page curves coincide in the thermodynamic limit.

The Gist

The Big Picture: Mixing Cards vs. Shuffling Quantum Magic

Imagine you have a deck of cards. Each card represents a tiny piece of information (a "qubit"). In the quantum world, these cards can be in a superposition of being both red and blue at the same time.

Physicists usually study how these cards get "entangled"—a spooky connection where the state of one card instantly tells you about the state of another, no matter how far apart they are. This usually happens when you use complex, magical quantum shuffles (random quantum gates).

This paper asks a different question: What happens if we only use classical shuffles? Imagine a machine that can only swap the positions of the cards (e.g., moving the Ace of Spades from position 1 to position 5) but cannot change the face of the card or create quantum magic. It's like a reversible classical computer.

The researchers wanted to know: Can a purely classical shuffling machine create deep quantum entanglement, or is it limited by how "quantum" the starting deck was?

---

Key Concept 1: The "Superposition" of the Starting Deck

The first major discovery is about the Initial State.

• The Analogy: Imagine you start with a deck where every card is clearly a "Red Ace." If you shuffle this deck classically, you just get a different order of Red Aces. Nothing interesting happens.
• The Twist: Now, imagine your starting deck is a "superposition." It's like a deck where every card is a blurry mix of Red and Blue.
• The Finding: The paper proves that the amount of quantum entanglement your classical shuffler can ever create is strictly limited by how "blurry" (superposed) your starting deck was.
  • If you start with a very "classical" deck (clear Red/Blue), the shuffler can't create much entanglement.
  • If you start with a "quantum" deck (very blurry), the shuffler can create a lot of entanglement.
  • The Metaphor: You can't bake a chocolate cake out of flour and water if you don't have any chocolate. The "quantumness" of the final cake is bounded by the "quantumness" of the ingredients you started with.

Key Concept 2: The "Page Curve" Race

The second part of the paper compares two different ways of mixing the deck:

1. The "Local" Mixer (The Circuit): Imagine a robot that only swaps two cards at a time, moving slowly across the deck. This is like a standard quantum circuit.
2. The "Global" Mixer (The Permutation): Imagine a wizard who can instantly swap any card with any other card in the entire deck all at once. This is a "global random permutation."

The Race:

• For Small Decks (Finite N): The two mixers behave differently. The "Local" mixer takes a specific amount of time to mix things up, and the "Global" mixer does it instantly. Their results (the "Page curves," which are graphs showing how entangled the system gets) look different.
• For Huge Decks (Thermodynamic Limit): Here is the surprise. As the deck gets infinitely large, the difference between the two mixers disappears. The "Local" robot eventually catches up to the "Global" wizard. They produce the exact same level of entanglement.

The Metaphor: Imagine two people trying to mix a giant pot of soup.

• Person A uses a tiny spoon (Local Circuit).
• Person B uses a giant industrial mixer (Global Permutation).
• In a small bowl, Person B is obviously faster and mixes it differently.
• But in an ocean-sized pot, if Person A keeps stirring long enough, the soup eventually becomes just as mixed as if Person B had used the industrial mixer. The "local" constraint fades away in the limit of infinite size.

Key Concept 3: Breaking the Rules (Adding Phases)

The researchers also asked: "What if we break the rules and add a little bit of 'quantum magic' (random phases) to our classical shuffler?"

• The Finding: If you add these random phases, the "Local" mixer and the "Global" mixer become identical even for small decks. The extra magic removes the barriers that kept them apart before.
• The Metaphor: It's like giving the person with the tiny spoon a little jetpack. Now they can keep up with the industrial mixer immediately, no matter how small the pot is.

Why Does This Matter?

This paper is important because it helps us understand the boundary between Classical and Quantum worlds.

1. It sets a limit: It tells us that classical circuits (which are easier to build and simulate) have a hard ceiling on how much quantum entanglement they can generate. They can't create "magic" out of thin air; they need "magic" ingredients to start with.
2. It reveals a hidden similarity: It shows that even though local interactions (swapping neighbors) seem very different from global interactions (swapping everyone), in the grand scheme of a massive system, they lead to the same chaotic, entangled outcome.

Summary in One Sentence

This paper shows that while classical shuffling machines are limited by how "quantum" their starting ingredients are, given enough time and a large enough system, they can eventually mimic the entanglement patterns of powerful, all-at-once quantum mixers.

Technical Summary

1. Problem Statement

The paper investigates the generation of quantum entanglement in many-body systems driven by random permutation circuits. Unlike standard random unitary circuits (which sample from the Haar measure or Clifford group), permutation circuits act classically on the computational basis by reshuffling basis states $|s\rangle \to |\pi(s)\rangle$.

The central questions addressed are:

1. Entanglement Bounds: What is the maximum entanglement that can be generated by classical (permutation) dynamics acting on an arbitrary initial quantum state?
2. Thermalization vs. Classicality: How does the entanglement dynamics of local random permutation circuits compare to global random permutations (the "infinite-depth" limit)? Do they converge to the same "Page curve" (entanglement entropy vs. subsystem size) in the thermodynamic limit?
3. Role of Quantumness: To what extent do "quantum" features (superpositions of classical states) influence entanglement generation in circuits that otherwise act classically?

2. Methodology

The authors employ a combination of rigorous analytical bounds, replica field theory, and exact numerical simulations.

• Models:

  • Global Permutations ($E_{GP}$): Sampling uniformly from the set of all $(2^N)!$ permutations acting on $N$ qubits.
  • Permutation Circuits ($E_{PC}$): Circuits composed of random 2-qubit permutation gates applied sequentially (depth $D$). The authors consider a continuous-time limit where gates are applied with rate $\lambda$.
  • Extensions: The study also considers circuits with random phases (automaton circuits) and $k$-local gates ($k \ge 3$).

• Analytical Tools:

  • Choi-Jamiolkowski Isomorphism: Used to map the density matrix evolution $\rho(t) \otimes \rho(t)$ to a state vector in a 4-replica Hilbert space.
  • Replica Space Dynamics: The ensemble-averaged purity $\mathbb{E}[\text{Tr}(\rho_A^2)]$ is mapped to the expectation value of a swap operator in the replica space. The evolution is governed by a Lindblad-like master equation: $\frac{d}{dt}\mathbb{E}[|\rho \otimes \rho\rangle] = -\mathcal{L}|\rho \otimes \rho\rangle$.
  • Symmetry Reduction:
    • For random product states, the authors exploit the fact that 2-qubit permutation gates belong to the Clifford group. This allows the dynamics to be projected onto a symmetric subspace of dimension $\binom{N+3}{3} \sim O(N^3)$, making exact numerical solutions feasible for large $N$ (up to $N=168$).
    • For global permutations, the lack of spatial structure simplifies the calculation to a combinatorial problem.

• Key Quantities:

  • Participation Entropy ($SPE_\alpha$): Measures the delocalization of the initial state in the computational basis.
  • Overlap ($z$): The overlap of the initial state with the "maximally antilocalized" state $|a_N\rangle = |+\rangle^{\otimes N}$.
  • Page Curves: The behavior of the averaged Rényi-2 entropy $S_2$ as a function of subsystem size $x = |A|/N$.

3. Key Contributions and Results

A. Tight Upper Bounds on Entanglement

The authors derive a rigorous upper bound on the Rényi-$\alpha$ entanglement entropy $S_\alpha(\rho_A(t))$ generated by any permutation circuit acting on an initial state $|\psi_0\rangle$:
$$S_\alpha(\rho_A(t)) \le \min\left[ |A|, \, SPE_\alpha(|\psi_0\rangle), \, (1-\alpha)^{-1}\log_2(z^{2\alpha}) \right]$$

• Significance: This result demonstrates that classical circuits cannot generate more entanglement than the "quantumness" of the initial state allows.
  • If the initial state is a computational basis state (classical), $SPE=0$ and $z=0$ (or specific values), resulting in zero entanglement.
  • If the initial state is a superposition (quantum), the entanglement is bounded by the degree of that superposition (measured by Participation Entropy and overlap with $|a_N\rangle$).
• Tightness: The authors prove analytically and verify numerically that for typical random permutation operators, this bound is saturated at leading order in $N$.

B. Comparison: Local Circuits vs. Global Permutations

The paper compares the late-time Page curves of:

1. Infinite-depth local circuits ($E_{PC}(\infty)$): 2-qubit gates applied randomly.
2. Global random permutations ($E_{GP}$): A single global shuffle.

• Finite $N$: The two ensembles yield different Page curves. The local circuit ensemble does not fully explore the space of global permutations due to locality constraints, even at infinite depth.
• Thermodynamic Limit ($N \to \infty$): The Page curves coincide. Both ensembles converge to the same functional form:
  $$\bar{s}_2(x) = \min\{x, \, s^{PE}_2, \, -2\log_2 \zeta_2\}$$
  where $s^{PE}_2$ is the intensive participation entropy and $\zeta_2$ relates to the overlap with the antilocalized state.
• Implication: This reveals an unexpected locality constraint on permutation operators. While they eventually thermalize to the same state as global permutations in the thermodynamic limit, the finite-size corrections differ significantly from standard Haar-random circuits.

C. Role of Clifford Structure and $k$-locality

• Clifford Gates: Since 2-qubit permutations are Clifford gates, the dynamics for random product states can be efficiently simulated. However, for arbitrary (non-stabilizer) initial states, the entanglement dynamics is hard to simulate, yet the authors show the bounds still hold.
• Breaking Clifford Symmetry ($k \ge 3$): When using 3-qubit (or higher) permutation gates, the gates are no longer Clifford. The authors show that for sufficiently deep circuits, these form permutation $t$-designs for all $t$. Consequently, the Page curves for local circuits and global permutations coincide for all finite $N$, not just in the thermodynamic limit.
• Random Phases: Adding random phases to the permutation gates (creating "automaton circuits") breaks the conservation of the antilocalized state. In this case, the bound simplifies to the Participation Entropy alone, and the local and global ensembles coincide for all $N$.

4. Significance and Implications

• Quantum vs. Classical Dynamics: The work provides a clear demarcation of how "quantumness" (superposition) is required to generate entanglement, even in systems driven by purely classical (permutation) dynamics. It quantifies the "resource" of initial state coherence needed for scrambling.
• Benchmarking Quantum Devices: Permutation circuits are easier to implement on current quantum hardware than continuous-time dynamics. The derived bounds and Page curves provide rigorous benchmarks for testing quantum advantage and entanglement generation in near-term devices.
• Thermodynamic Limits: The finding that local and global permutation circuits converge to the same Page curve only in the thermodynamic limit highlights subtle finite-size effects in many-body quantum chaos that differ from Haar-random models.
• Exact Solvability: The derivation of exact differential equations for the purity dynamics in the replica space offers a new analytical framework for studying non-Haar random circuits, specifically those with discrete symmetries.

In summary, the paper establishes that while classical permutation circuits are limited by the initial state's quantum coherence, they exhibit rich entanglement dynamics that, in the thermodynamic limit, mimic the scrambling properties of fully random unitary circuits, subject to specific finite-size constraints.