Stochastic Krylov Dynamics: Revisiting Operator Growth in Open Quantum Systems

This paper demonstrates that while operator growth in closed quantum systems follows deterministic Hamiltonian flow in an emergent phase space, coupling to an environment transforms this dynamics into a stochastic process where dissipation induces diffusion and ultimately destroys the hyperbolic mechanism responsible for exponential complexity growth.

The Gist

The Big Picture: How Things Get Complicated

Imagine you have a simple toy, like a single Lego brick. In a perfect, isolated world (a "closed system"), if you shake the box, that brick might bounce around and eventually attach to other bricks, forming a giant, complex structure. In physics, we call this operator growth. It's how a simple piece of information spreads out and becomes entangled with the rest of the system, making it impossible to track where it started.

In the past, physicists found a beautiful way to describe this: they imagined the Lego brick moving along a straight, infinite track. The rules of the track were fixed, and the brick moved in a predictable, deterministic way. If the track had a certain shape (like a steep hill), the brick would speed up exponentially, representing chaos.

The Problem: Real life isn't a perfect, isolated box. Our toys are in a room with wind, dust, and people bumping into them. In physics, this is an open system interacting with an environment. The old rules didn't work anymore because the environment introduces noise, friction, and randomness.

The Solution: This paper builds a new map for how things get complicated when the environment is involved. It turns out that instead of a smooth, predictable train ride, the journey becomes a stochastic (random) walk, like a drunk person trying to walk down a hallway while being pushed by the wind.

---

Key Concepts Explained with Analogies

1. The "Krylov Chain" (The Infinite Track)

Think of the "Krylov space" as a long, one-dimensional hallway with numbered rooms (0, 1, 2, 3...).

• Room 0: Your simple starting Lego brick.
• Room 100: A massive, complex structure made of thousands of bricks.
• The Goal: To see how fast your brick travels from Room 0 to Room 100.

In a perfect world, the brick moves down this hallway according to strict laws of physics. The paper shows that even in a messy world, this hallway still exists, but the rules of movement change.

2. The Two Ways the Environment Messes Things Up

The paper identifies two main ways the environment (the "wind and dust") changes the journey.

Scenario A: The "Drunk Walker" (Pure Dephasing)
Imagine your brick is trying to run down the hallway, but a gusty wind is blowing it sideways.

• The Old View: The brick runs straight down the center.
• The New View: The brick still tries to run down the center, but the wind pushes it left and right randomly.
• The Result: The brick still moves forward, but its path is wobbly. Sometimes it speeds up, sometimes it slows down.
• The Analogy: It's like trying to walk a tightrope while a storm blows. You are still moving forward, but your path is no longer a straight line; it's a random walk. The paper calls this "Noisy Hyperbolic Flow." The chaos is still there, but it's "fuzzy" and less predictable.

Scenario B: The "Sponge" (Absorptive Potential)
Now, imagine the hallway isn't just windy; the floor is made of a giant sponge.

• The Mechanism: The further you go down the hallway (the more complex your structure gets), the more the sponge sucks you in.
• The Result: Simple structures (near Room 0) survive easily. Complex structures (near Room 100) get "eaten" by the sponge before they can fully form.
• The Analogy: It's like trying to build a sandcastle on a beach while the tide is coming in. You can build a small castle, but if you try to build a huge one, the water washes it away before it's finished.
• The Outcome: The complexity stops growing and hits a "ceiling." It saturates. The system never gets fully chaotic because the environment kills the complexity before it can spread.

3. The "Race" (Scrambling vs. Dissipation)

The paper frames the whole situation as a race between two forces:

1. The Runner (Chaos): The natural tendency of the system to spread information and get complex (running down the hallway).
2. The Obstacle (Dissipation): The environment trying to stop it (the wind or the sponge).

• If the Runner is faster: The system "scrambles" (becomes chaotic and complex) before the environment can stop it.
• If the Obstacle is stronger: The system gets stuck or localized. The complexity never fully develops.

This creates a Phase Transition. Just like water turning to ice, a quantum system can switch from "Chaos Mode" to "Frozen/Stuck Mode" depending on how strong the environment is.

---

Why Does This Matter?

1. It's not just "broken" chaos; it's a new kind of chaos.
Previously, scientists thought open systems were just closed systems with some "noise" added on top. This paper shows that the noise fundamentally changes the geometry of the problem. The smooth, predictable path turns into a probabilistic cloud of possibilities.

2. It explains why some things don't get complicated.
In the real world, we see systems that stay simple or get stuck. This paper gives us the mathematical tools to predict when a system will stay simple (because the environment is too strong) and when it will get chaotic.

3. A New Tool for the Future.
The authors developed a new mathematical "lens" (called the Schwinger-Keldysh formulation) that allows us to calculate these messy, real-world scenarios. It's like upgrading from a black-and-white map to a 3D GPS that accounts for traffic, weather, and roadblocks.

Summary in One Sentence

This paper shows that when quantum systems interact with their environment, the neat, predictable path of becoming complex turns into a random, noisy journey where the environment acts like a sponge or a windstorm, either slowing down the chaos or stopping it entirely, depending on who wins the race.

Technical Summary

1. Problem Statement

The paper addresses the challenge of characterizing operator growth in open quantum systems.

• Context: In closed quantum systems, operator growth is well-understood via Krylov complexity. Using the Lanczos algorithm, the Liouvillian (commutator with the Hamiltonian) is mapped to a tridiagonal "Krylov chain." This allows for a geometric interpretation where operator growth corresponds to a particle moving in an emergent phase space $(n, p)$, governed by a classical Hamiltonian flow. Exponential growth (scrambling) arises from hyperbolic instability in this phase space.
• The Gap: Most physical systems are open, interacting with environments via dissipation and noise (described by Lindblad dynamics).
  • The standard Krylov construction relies on the Liouvillian being anti-Hermitian (unitary evolution). In open systems, the Lindbladian is non-Hermitian, breaking the standard Lanczos recursion and the unitary preservation of operator norms.
  • It is unclear how the geometric phase-space picture of operator growth survives dissipation. Does dissipation merely renormalize the growth rate, or does it fundamentally alter the dynamical mechanism (e.g., destroying chaos)?

2. Methodology

The authors extend the Schwinger–Keldysh (SK) path integral formulation of Krylov complexity (previously developed for closed systems) to open quantum systems.

• Full Counting Statistics (FCS): They define the generating functional for the Krylov position operator $\hat{n}$ as $Z(\chi, t) = \langle e^{i\chi \hat{n}(t)} \rangle$.
• Contour Formulation: They construct a real-time SK path integral on a closed time contour (forward $+$ and backward $-$ branches).
  • Closed Systems: The SK action reduces to a classical Hamiltonian flow in the limit of small quantum fluctuations.
  • Open Systems: They incorporate the Lindblad master equation directly into the SK formalism. The dissipative effects appear as an influence functional coupling the forward and backward branches.
• Two Distinct Approaches:
  1. Stochastic Approach (Sections 4-5): They model pure dephasing where the jump operator is proportional to the Krylov position ($L \propto \hat{n}$). This leads to a non-unitary evolution that can be mapped to a stochastic differential equation (Langevin dynamics) in the emergent phase space.
  2. Non-Hermitian Projection Approach (Section 6): They project the full Lindbladian onto the Krylov basis of the closed system, deriving an effective non-Hermitian tight-binding Hamiltonian. They then construct the SK path integral for this non-Hermitian system, treating dissipation as an absorptive potential.

3. Key Contributions and Results

A. Stochastic Dynamics in Emergent Phase Space (Pure Dephasing)

For the case of pure dephasing ($L = \sqrt{\kappa} \hat{n}$), the authors derive the effective SK action:
$$S_{SK} = S_{cl} + i \frac{\kappa}{2} \int dt \, n_q^2$$
where $n_q$ is the "quantum" field (difference between forward and backward trajectories).

• Result: Integrating out the quantum fields transforms the deterministic Hamiltonian equations of motion into Langevin equations:
  $$\dot{n} = \partial_p H_{eff}, \quad \dot{p} = -\partial_n H_{eff} + \eta(t)$$
  where $\eta(t)$ is Gaussian white noise with strength $\kappa$.
• Physical Interpretation: The "hyperbolic instability" responsible for exponential growth in closed systems becomes a noisy hyperbolic flow.
  • The unstable manifold (responsible for exponential growth $e^{2\alpha t}$) is broadened into a stochastic tube of width $\delta p_{rms} \sim \sqrt{\kappa/\alpha}$.
  • Renormalized Growth: The typical Lyapunov exponent is renormalized from $2\alpha$ to $\lambda_{typ} = 2\alpha - \kappa/4$.
  • Regimes:
    • $\kappa \ll \alpha$: Weakly renormalized chaos.
    • $\kappa \sim \alpha$: Strong fluctuations, intermittent growth.
    • $\kappa \gg \alpha$: The unstable manifold is washed out; exponential growth is destroyed.

B. Non-Hermitian Projection and Absorptive Potentials

In the alternative approach, the Lindblad dynamics is projected onto the Krylov basis, yielding an effective non-Hermitian Hamiltonian:
$$\hat{H}_{NH} = \hat{H}_{TB} - i \gamma \hat{d}(\hat{n})$$
where $\hat{d}(\hat{n})$ acts as an imaginary potential increasing with Krylov depth (complexity).

• Result: The SK path integral remains deterministic at the saddle-point level, but trajectories are weighted by an absorptive factor $\exp(-\int \gamma d(n) dt)$.
• Physical Interpretation:
  • The system obeys a "principle of least decay." While the classical equations of motion still allow for exponentially growing trajectories, these paths are exponentially suppressed in the path integral due to the high "cost" (decay) of reaching large Krylov depths.
  • Saturation: Long-time dynamics are dominated by edge-localized trajectories (low complexity). Consequently, Krylov complexity does not grow indefinitely but saturates to a finite value $K_\infty$.
  • Scrambling vs. Dissipation: A dynamical phase transition occurs based on the competition between the scrambling time and the dissipation time (localization length $\xi$). Scrambling only occurs if the Krylov dimension $D_K < \xi$.

4. Significance and Implications

• Unified Framework: The paper provides a unified geometric and stochastic framework for operator growth in open systems, bridging the gap between unitary chaos and dissipative dynamics.
• Dissipation as a Relevant Perturbation: It identifies dissipation not just as a perturbation that slows down growth, but as a mechanism that fundamentally changes the nature of the dynamics from deterministic Hamiltonian flow to stochastic dynamics or spectral selection.
• New Universality Classes: The work suggests that open-system operator growth defines new universality classes distinct from closed systems, governed by the interplay of the growth rate ($\alpha$) and noise/dissipation strength ($\kappa$ or $\gamma$).
• Experimental Relevance: The results offer concrete predictions for experiments involving monitored quantum systems, measurement-induced phase transitions, and thermalization in open quantum matter, specifically regarding how information scrambling survives or is destroyed by environmental coupling.
• Future Directions: The authors outline paths for extending this to thermal baths (incorporating fluctuation-dissipation relations) and spatially extended systems, potentially linking Krylov complexity to quantum thermodynamics.

In summary, the paper demonstrates that operator growth in open systems is a problem of stochastic dynamics in an emergent phase space, where the competition between chaotic instability and environmental noise/dissipation determines whether information scrambling survives.