Algebraic Tomography of Non-Hermitian Floquet Systems from Observable Traces

This paper formulates a framework for algebraic tomography of finite-dimensional non-Hermitian Floquet systems that reconstructs spectral data from observable trace sequences using characteristic polynomial constraints and resolvent methods, while clarifying identifiability limits and demonstrating applications in driven qutrits and SSH chains.

The Gist

Imagine you are trying to figure out the inner workings of a mysterious, ticking clock (the quantum system) that repeats its motion every second. You can't open the clock to see the gears inside. All you have is a single, small window on the side where you can peek and see a shadow moving back and forth (the observable).

This paper, titled "Algebraic Tomography of Non-Hermitian Floquet Systems from Observable Traces," proposes a new, highly mathematical way to reconstruct the entire clock mechanism just by watching that shadow move over time.

Here is the breakdown of their ideas using simple analogies:

1. The Problem: The "Black Box" Clock

In physics, many systems (like atoms or circuits) are driven by a repeating rhythm. Physicists call the "one full turn" of this rhythm the Monodromy Matrix. It's the master blueprint of the system.

• The Catch: You usually can't see the blueprint. You can only measure specific things, like "how much energy is in the top part of the clock?" or "how bright is the light?"
• The Old Way: Usually, scientists try to guess the blueprint by fitting a curve to the data, like guessing the shape of a hidden object by tracing its shadow. This often leads to errors or requires huge amounts of data.

2. The New Idea: The "Skeleton vs. The Costume"

The authors realized that the shadow you see isn't just random noise; it's strictly constrained by the math of the clock's gears. They call their method Floquet Algebraic Tomography.

They split the problem into two parts:

• The Skeleton (The Gears): This is the core structure of the clock. It's the same no matter what you look at. It determines the fundamental "notes" or frequencies the clock can play.
• The Costume (The Dressing): This is how your specific window (the observable) sees the gears. If you look through a red filter, the shadow looks red. If you look through a blue filter, it looks blue. The "costume" changes based on where you stand, but the "skeleton" underneath stays the same.

The Analogy: Imagine a puppet show.

• The Skeleton is the puppeteer's hand movements (the true physics).
• The Costume is the puppet's outfit.
• The Trace is the shadow the puppet casts on the wall.
• The authors' method allows you to figure out exactly how the puppeteer's hand is moving (the skeleton) just by analyzing the shadow, even if the puppet is wearing a different costume (a different measurement tool) every time.

3. How They Do It: The "Magic Recurrence"

Instead of guessing, they use a mathematical rule called the Cayley-Hamilton theorem. Think of this as a "magic recipe."

• If you watch the shadow for just a few seconds, this recipe tells you exactly how long the sequence of movements will repeat.
• It acts like a sieve. It separates the Skeleton (the universal rules of the clock) from the Costume (the specific way your measurement sees it).
• They use a tool called a Hankel Matrix (think of it as a giant spreadsheet of the shadow's history) to organize this data. By looking at the patterns in the spreadsheet, they can mathematically "realize" or rebuild a copy of the clock's master blueprint.

4. The Limits: What You Can't See

The paper also honestly discusses what happens if your window is too small or if the clock has a secret symmetry.

• The Invisible Sector: Imagine the clock has a hidden compartment that your window can never see. No matter how long you watch, you can't know what's in that compartment. The math proves that if your "window" (observable) is too limited, you will only ever see a "shadow version" of the clock, not the real thing.
• Micromotion (The Magic Trick): The authors show that if you can slightly shift when you start watching (a concept called micromotion), you can change the angle of your window. This is like moving your head slightly to see around a corner. It helps you see more of the clock's gears.
• The Symmetry Wall: However, if the clock has a perfect symmetry (like a perfectly balanced wheel), even moving your head won't help. Some parts of the clock will remain permanently invisible because the symmetry hides them mathematically.

5. Two Real-World Tests

To prove their method works, they tested it on two scenarios:

• Test 1: The Leaky Qubit (A Quantum Computer Bit):
  They simulated a superconducting qubit (a type of quantum bit) that sometimes "leaks" energy into a third, unwanted level.

  • Result: When the qubit was isolated, their method saw only a tiny, one-dimensional shadow. But when "leakage" turned on, the shadow suddenly expanded to fill the whole space. Their math successfully detected this "leakage" by noticing the shadow grew larger, proving the system was more complex than just a simple two-level bit.

• Test 2: The SSH Chain (A Line of Atoms):
  They simulated a chain of atoms where particles hop from one to another, but the hopping is "non-reciprocal" (it's easier to hop left than right).

  • Result: They showed that depending on which atom you measure, you see a completely different story. Sometimes the shadow shows a "winding" pattern (a topological feature), and sometimes it looks flat. Their method explained why this happened: the "costume" (the specific atom you chose to measure) was filtering out the "skeleton's" true shape.

Summary

This paper doesn't invent a new physical machine; it invents a new mathematical lens.
It tells physicists: "Don't just try to fit a curve to your data. Use the strict rules of algebra to separate the universal truth of your system from the bias of your measurement tool."

It provides a rigorous way to say: "Here is the part of the system I can see, and here is the part that is mathematically invisible to my current tools." This helps researchers understand exactly how much of a quantum system they are actually observing and how much is hidden in the shadows.

Technical Summary

Technical Summary: Algebraic Tomography of Non-Hermitian Floquet Systems from Observable Traces

Problem Statement
The central problem addressed is the inverse problem of reconstructing the properties of a finite-dimensional, non-Hermitian Floquet system from limited observational data. In such systems, the fundamental object is the monodromy matrix $M \in GL(N, \mathbb{C})$, which governs the stroboscopic evolution over one period $T$. However, $M$ is typically not directly accessible. Instead, experimental access is restricted to time-discrete response sequences generated by a fixed observable $O$, defined as the Observable Trace Sequence (OTS):
$$\zeta^{(O)}_n := \text{Tr}(O M^n), \quad n \in \mathbb{N}_0.$$
While such sequences resemble generic exponential signals, they are strictly constrained by the exact algebraic structure of the underlying finite-dimensional matrix $M$. The challenge is to determine what aspects of the monodromy matrix (the "spectral skeleton") and its dynamics can be uniquely reconstructed from these traces, and what aspects remain "invisible" due to the specific choice of observable, system symmetries, or finite-size effects. This is particularly delicate in non-Hermitian regimes where exceptional points (EPs), non-Hermitian skin effects, and observable-dependent cancellations coexist.

Methodology
The paper proposes a framework termed Floquet algebraic tomography, which treats the reconstruction not as a generic exponential fitting problem (e.g., Prony or Matrix Pencil methods) but as a finite-dimensional algebraic reconstruction problem constrained by the Cayley–Hamilton (CH) theorem.

1. Analytic Decomposition (Skeleton vs. Dressing):
   The framework introduces three analytic components derived from the OTS:

   • Observable Resolvent (ORS): $Z^{(O)}(z) = \sum \zeta^{(O)}_n z^{n-1}$. Crucially, this is decomposed as $Z^{(O)}(z) = \frac{N^{(O)}(z)}{\Delta(z)}$, where the denominator $\Delta(z) = \det(I_N - zM)$ is the characteristic polynomial common to all observables (the "skeleton"), and the numerator $N^{(O)}(z)$ carries the observable-dependent "dressing."
   • Observable Spectral Determinant (OSD): $h^{(O)}(z) = \exp[-\text{Tr}(O \log(I_N - zM))]$.
   • Observable Dirichlet Spectral Data (ODSD): $F^{(O)}(p)$, providing a polylogarithmic readout of the spectral data.
     This separation allows the framework to distinguish between the intrinsic spectral geometry of $M$ and the channel-specific visibility imposed by $O$.

2. Realization-Theoretic Reconstruction:

   • Single Observable: Using the CH recurrence relation induced by the OTS, the effective recurrence order and characteristic coefficients $\{e_a\}$ are recovered via Hankel matrix analysis. This yields the spectrum $\{\lambda_j\}$ and observable weights $\{c^{(O)}_j\}$.
   • Multi-Observable Extension: By constructing block-Hankel matrices from multiple observables $\{O_\ell\}$, the framework employs a Ho–Kalman-like realization to reconstruct a matrix $M_{\text{rel}}$ similar to the true monodromy $M$ ($M_{\text{rel}} = S M S^{-1}$). This recovers the similarity-invariant spectral skeleton.
   • Liouville-Space Mapping: The framework extends to fundamental-type ($\text{Tr}(O M^n \Omega)$) and adjoint-type ($\text{Tr}(O M^n \Omega M^{-n})$) sequences, connecting to quantum process tomography and gate-set tomography. This allows for the resolution of degeneracies (Diabolic Points) by varying initial states $\Omega$.

3. Algebraic Identifiability and Micromotion:
   The paper analyzes the limits of identifiability using operator algebra. If the accessible observables generate a proper subalgebra $\mathcal{O} \subsetneq \text{End}(\mathcal{H})$, the full matrix $M$ is not identifiable. Instead, the data determines a canonical visible representative $M_{\text{vis}} = \mathcal{E}_{\mathcal{O}''}(M)$, the projection of $M$ onto the bicommutant $\mathcal{O}''$, plus a trace-invisible remainder.

   • Micromotion: Shifting the reference time $t$ generates a time-shifted observable family $O(t) = U(t,0)^{-1} O U(t,0)$. This dynamically enlarges the observable algebra, potentially expanding the visible operator space.
   • Symmetry Constraints: The paper proves that exact commuting symmetries impose a non-removable algebraic deficiency; even with micromotion, sectors commuting with the symmetry remain invisible if the initial observable algebra does not break them.

Key Results and Examples
The framework is validated through two numerical examples:

1. Driven Transmon Qutrit (DTQ3):

   • A $GL(3, \mathbb{C})$ system modeling a superconducting qubit with coherent leakage to a third level.
   • Reconstruction: A single observable channel successfully reconstructs the full spectral skeleton to machine precision.
   • Visibility: The ODSD analysis reveals that the visible phase response is a filtered projection of the global determinant phase, subject to destructive interference from observable dressing.
   • Leakage Detection: The sampled observable dimension $D_{\text{obs}}$ (rank of the micromotion-expanded map) remains 1 in the isolated qubit limit but expands to $N^2=9$ when coherent leakage is active, algebraically detecting the expansion of the accessible operator space.

2. Finite Non-Hermitian Floquet SSH Chain:

   • A finite-size, non-reciprocal SSH chain exhibiting exceptional points (EPs) and non-Hermitian skin effects.
   • EP Geometry: The framework distinguishes between the branch-sensitive spectral skeleton (inherited from the Bloch-sector EP) and the observable-dependent visible response. Local or staggered observables can suppress topological winding signals via destructive interference, while collective probes retain partial visibility.
   • Disorder and Symmetry: The sampled observable dimension $D_{\text{obs}}$ is shown to be controlled jointly by exact symmetries, finite-size degeneracies, and disorder. Bond disorder lifts accidental linear dependencies, increasing $D_{\text{obs}}$, whereas on-site disorder does not necessarily guarantee saturation of the algebraic bound.

Significance and Claims
The paper positions itself as a structural and inverse-problem contribution rather than a proposal for new physical phenomena. Its primary significance lies in:

• Formalizing the Inverse Problem: Establishing that Floquet OTS data constitutes a constrained algebraic problem rather than a generic signal processing task, leveraging exact CH identities.
• Skeleton/Dressing Separation: Providing a rigorous analytic tool (ORS/OSD/ODSD) to separate the universal spectral geometry of the monodromy matrix from the channel-specific artifacts introduced by the measurement probe.
• Quantifying Visibility: Defining and calculating the "observable dimension" and "algebraic deficiency" to precisely quantify what parts of the dynamics are reconstructible versus what remains invisible due to symmetry or restricted observables.
• Bridging Theory and Experiment: Connecting abstract algebraic realization theory (Hankel matrices, commutants) to practical quantum tomography settings (process tomography, gate-set tomography, and stroboscopic spectroscopy), offering a language to interpret why certain topological or spectral features appear, vanish, or distort in specific experimental channels.

The authors emphasize that this framework clarifies the limits of identifiability: even with perfect data, the full monodromy matrix may remain partially invisible if the observable algebra is insufficient, and that micromotion can only enlarge visibility up to the bounds imposed by exact symmetries.