Phases of Floquet code under local decoherence

This paper investigates the robustness of Floquet codes under local decoherence by deriving a 3D statistical mechanics model for their maximum likelihood decoder, identifying specific Pauli channels with decoupled thresholds, and proposing a diagnostic that distinguishes the code's anyon automorphism phase from the toric code via a phase transition at the decoherence threshold.

The Gist

Imagine you have a magical, self-correcting library where books (quantum information) are constantly being shuffled around by a librarian who follows a strict, repeating schedule. This library is called the Floquet Code.

In a normal library, if you want to keep a book safe, you lock it in a vault. But in this magical library, the "vault" isn't a static room; it's a dynamic dance. Every few minutes, the librarian performs a specific set of checks (measurements) that rearrange the books. The unique twist? After one full cycle of these checks, the books don't just return to their original spots; they swap roles. A book that was acting like a "Red" book suddenly starts acting like a "Blue" book. This role-swapping is called anyon automorphism.

The authors of this paper asked a crucial question: What happens if the library is noisy? What if the librarian is slightly distracted, or the books get bumped by random vibrations (decoherence)? Does the library still work, and does the magical role-swapping survive?

Here is the breakdown of their findings using simple analogies:

1. The Problem: Noise in the System

Think of "decoherence" as static on a radio or dust settling on a lens. In a quantum computer, this noise scrambles the information. Usually, if you have too much noise, the information is lost forever. The researchers wanted to find the "threshold": the exact amount of noise the library can tolerate before it collapses.

2. The Detective Work: Finding the Threshold

To find this limit, the authors acted like detectives trying to reconstruct a crime scene.

• The Clues: They looked at "syndromes," which are like footprints left behind by errors.
• The Map: They realized that figuring out the best way to fix the errors (decoding) is mathematically equivalent to solving a complex 3D puzzle.
• The Simplification: They found a special type of "simple error" (like a specific kind of dust) where this 3D puzzle breaks apart into a stack of 2D puzzles. This made it much easier to solve.
• The Result: They calculated the exact tipping point. If the noise is below 1.19%, the library can still recover the information perfectly. If it's above that, the information is lost.

3. The Magical Feature: The Role-Swap (Anyon Automorphism)

This is the most exciting part. In a standard quantum library (called the Toric Code), the books stay in their roles. If you check the library, the "Red" book is always the "Red" book.

But in the Floquet Code, the library has a personality. After every full cycle of checks, the "Red" book becomes the "Blue" book, and the "Blue" becomes the "Red."

• The Test: The authors created a special test (using something called "quantum relative entropy") to see if this role-swapping was still happening even when the library was dusty.
• The Finding: Below the noise threshold, the magic survives. The books still swap roles perfectly. The library knows it's a Floquet library, not a standard one.
• The Contrast: If you tried this same test on a standard Toric Code library, the books would never swap roles. The test would give a different result, proving that the Floquet Code is a completely different "species" of quantum memory.

4. The Verdict

The paper concludes that the Floquet Code is robust.

• Below the threshold: The library is healthy. It can fix errors, and its unique "role-swapping" feature remains intact. It is a distinct, stable phase of matter.
• Above the threshold: The noise is too loud. The library forgets the books, and the magical role-swapping stops. It collapses into a "trivial" state where no information is stored.

Summary Analogy

Imagine a dance troupe performing a complex routine where dancers constantly switch partners and costumes.

• The Noise: Random people bumping into the dancers.
• The Threshold: The point where the bumps are so frequent that the dancers can't keep the routine together.
• The Discovery: The authors proved that as long as the bumps are rare (under 1.19%), the troupe can not only keep dancing but also keep their unique "switching partners" routine. This proves they are a special kind of troupe, distinct from a group that just stands in place.

The paper does not claim this can be used to build a specific product today or cure diseases. It strictly proves that this specific type of quantum memory has a mathematically defined "safe zone" where it works and retains its unique, dynamic properties.

Technical Summary

Technical Summary: Phases of Floquet Code under Local Decoherence

Problem Statement
The Hastings-Haah Floquet code is a dynamical quantum memory characterized by a periodically evolving logical space and a unique feature: an anyon automorphism (specifically an $e$-$m$ exchange) occurring after each measurement period. While it is established that this code can robustly encode quantum information in the presence of local decoherence, the intrinsic error threshold for the Floquet code under perfect measurements has remained undetermined. Furthermore, it is unclear whether the defining feature of the code—the anyon automorphism—remains well-defined below the threshold or if the code transitions into a standard topological phase (like the toric code) or a trivial phase. Unlike static topological codes where stabilizers are measured directly, the Floquet code's stabilizers are inferred only after two consecutive rounds of measurements, leading to time-correlated syndrome changes that complicate the derivation of decoding thresholds and phase diagnostics.

Methodology
The authors employ two complementary approaches to determine the intrinsic threshold and characterize the phases of the decohered Floquet code:

1. Maximum Likelihood Decoding (MLD) and Statistical Mechanics Mapping:
   The authors derive the statistical mechanics model governing the maximum likelihood decoder for the Floquet code under local Pauli decoherence. They identify a specific class of "simple" two-qubit Pauli errors (e.g., $E^Y_R, E^Z_R$, etc.) where syndromes can be read out after a single round of measurements. For these errors, the authors demonstrate that the generally coupled 3D statistical mechanics model (arising from time-correlated syndromes) decouples into a stack of independent 2D Random Bond Ising Models (RBIM) on a honeycomb lattice. The decoding threshold is then identified with the ferromagnetic-to-paramagnetic phase transition point of these RBIMs.

2. Information-Theoretic Diagnostics:
   To probe the phase structure and the persistence of the anyon automorphism, the authors propose diagnostics based on quantum relative entropy and coherent information.

   • $e$-$m$ Automorphism Probe: They define a quantity $D_{em}$ based on the quantum relative entropy between two states: one initialized in an $m$-logical eigenstate and the other in a maximally mixed state, both evolved through the Floquet sequence. The second state is projected onto the $e$-logical eigenstate that the first state would evolve into in the absence of decoherence.
   • Coherent Information: They compute the coherent information $I_c$ to distinguish the encoding phase from the trivial phase.
   • Statistical Mapping: Using the stabilizer expansion of the decohered density matrix, they map both the Rényi relative entropy $D^{(n)}_{em}$ and the Rényi coherent information $I^{(n)}_c$ to partition functions of $(n-1)$-flavor Ising models. This allows for an analytical study of the phase transitions.

Key Contributions and Results

• Derivation of the Intrinsic Threshold: The authors establish that for the class of simple errors, the Floquet code undergoes a phase transition at a critical error rate $p_c \approx 0.0119$. This value corresponds to the critical point of the 2D RBIM on the Nishimori line.
• Decoupling of the 3D Model: A central technical result is the identification of error channels where the 3D statistical mechanics model for the decoder decouples into 2D models in the time direction. This simplifies the analysis significantly compared to the general case where time-direction coupling persists.
• Robustness of Anyon Automorphism: The authors analytically demonstrate that below the threshold $p_c$, the Floquet code exhibits a robust phase characterized by the $e$-$m$ anyon automorphism. In this phase, the relative entropy probe $D^{(n)}_{em}$ vanishes ($D^{(n)}_{em} = 0$), indicating that the logical state evolution preserves the automorphism property.
• Phase Distinction: The diagnostics distinguish three distinct phases:
  1. Floquet Code Phase ($p < p_c$): Characterized by $I_c = 2 \log 2$ (maximum coherent information) and $D^{(n)}_{em} = 0$. The logical space undergoes non-trivial $e$-$m$ automorphism.
  2. Toric Code Phase: A phase retaining information but lacking $e$-$m$ automorphism (e.g., if specific measurements are omitted). Here, $I_c = 2 \log 2$ but $D^{(n)}_{em} = (\log 2)/(n-1)$.
  3. Trivial Phase ($p > p_c$): Characterized by non-positive coherent information ($I_c \leq 0$) and $D^{(n)}_{em} = \log 2$, indicating a loss of quantum information.
• Simultaneous Transition: The paper shows that the transition in the coherent information (marking the loss of information) and the transition in the automorphism probe occur simultaneously at the same threshold $p_c$.
• Optimality of MLD: By showing that the threshold derived from the coherent information matches the threshold derived from the maximum likelihood decoder (via the Kramers-Wannier duality between the $(n-1)$-flavor Ising model and the RBIM), the authors confirm that the maximum likelihood decoder is asymptotically optimal for the Floquet code under these conditions.

Significance
The paper establishes the intrinsic error threshold for the Hastings-Haah Floquet code under perfect measurements and local Pauli decoherence. Crucially, it proves that the code's defining dynamical feature—the anyon automorphism—is a robust property of the encoding phase, persisting up to the same threshold where quantum information is lost. This distinguishes the Floquet code as a separate dynamical phase from the static toric code, even in the presence of decoherence. The work provides a rigorous statistical mechanics framework for analyzing dynamical quantum memories and offers a concrete diagnostic tool (based on relative entropy) to experimentally verify the presence of anyon automorphism in noisy intermediate-scale quantum devices.