🔬 physics

Hamiltonian dynamics from pure dissipation

This paper demonstrates that internal Hamiltonian dynamics can be effectively simulated using only external pure dissipation (Lindbladians without a coherent part) with optimal time scaling, establishing fundamental limits on the decoherence cost required to mimic reversible evolution and revealing significant implications for quantum complexity, simulation efficiency, and dynamical control.

Original authors: Daniel Stilck França, Zhong-Xia Shang

Published 2026-07-31

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniel Stilck França, Zhong-Xia Shang

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Invisible Whip and the Spinning Top

Imagine a world where everything that moves is either a perfectly isolated dancer or a clumsy person bumping into a crowded room. In the realm of quantum physics, this is the difference between a "closed" system and an "open" one. A closed system is like a dancer in a soundproof, glass-walled studio: they spin and leap with perfect precision, following a strict internal rhythm (called a Hamiltonian) without ever losing a step or leaking energy. This is the ideal, reversible motion that powers the most advanced quantum computers.

On the other hand, an "open" system is like that same dancer trying to perform in a bustling marketplace. They are constantly bumped by passersby (the environment), causing them to stumble, lose their balance, and eventually stop. In physics, this is called "dissipation" or "decoherence," and it's usually seen as the enemy of quantum magic because it destroys information. For decades, scientists believed these two worlds were fundamentally incompatible: you couldn't get the perfect, reversible spin of the dancer just by having them bump into the crowd. But what if the crowd could actually teach the dancer a new way to spin? What if the very thing that usually breaks a quantum system could be used to build one?

The Great Imitation: Faking a Spin with a Whip

In this new work, researchers Daniel Stilck França and Zhong-Xia Shang have discovered a surprising trick: you can "fake" the perfect, reversible spin of a closed quantum system using only the messy, irreversible bumps of an open system. They call this "Hamiltonian dynamics from pure dissipation."

To understand their discovery, picture a spinning top. Normally, a top spins because you gave it an initial twist (an internal force). If you stop pushing it, friction and air resistance (dissipation) will eventually make it wobble and fall. The researchers asked: Can we make a top spin perfectly without ever giving it an initial twist, by only hitting it with a whip from the outside?

Their answer is a resounding yes. They proved that if you hit the top with a very specific, rhythmic series of tiny whips (which they call "jump operators"), the top will start to rotate just as if it had an internal motor. The catch is that every time you whip the top to make it spin, you also accidentally give it a tiny, unwanted wobble (radial disturbance). However, by timing these whips perfectly, the spinning motion becomes the dominant effect, and the wobble becomes negligible.

The paper shows that to mimic a perfect spin for a time $t$ with a tiny error $\epsilon$ , you need to run this "whipping" process for a total time of roughly $O(t^2/\epsilon)$ . This means the cost of faking the spin grows with the square of the time you want to spin and the inverse of how perfect you want it to be. Crucially, this is efficient; you don't need an impossible amount of energy or time. It's like realizing that while you can't stop the wind from blowing, you can use the wind to power a sailboat just as effectively as an engine, provided you adjust your sails correctly.

The Geometry of the Trade-Off

The researchers didn't just find a way to do this; they proved it's the best you can possibly do. They used a geometric argument to show that you cannot cheat the laws of physics here. Imagine trying to rotate a ball on a string. To make it spin fast (like a Hamiltonian), you have to pull on the string. But pulling on the string inevitably pulls the ball inward (like dissipation).

The paper proves that there is a strict trade-off: you cannot increase the speed of the spin without increasing the inward pull. If you try to spin the top faster, the "wobble" (decoherence) gets worse. The $O(t^2/\epsilon)$ time cost is the mathematical price you must pay to overcome this inevitable wobble. It's the "decoherence cost" of catching up to the speed of a perfect, internal spin using only external pushes. The authors show that in the worst-case scenario, you cannot do better than this scaling; it is a fundamental limit, not just a limitation of current technology.

From Theory to Practice: The "QDRIFT" Whip

Beyond the theory, the authors show how to turn this idea into a practical tool. They propose a method called "Autonomous QDRIFT." In standard quantum computing, to simulate a complex system, you often have to randomly choose which "move" to make next, like a chef randomly picking spices. This usually requires a complex computer to keep track of the random choices.

In this new method, the randomness comes naturally from the environment itself. Imagine a machine that randomly "clicks" (like a Geiger counter) and, based on those clicks, automatically applies the right "whip" to the quantum system. No external computer is needed to tell the system what to do; the noise of the environment is the controller. This is especially useful for machines where it's hard to perform precise, controlled moves but easy to let the system interact with a noisy environment.

Furthermore, the paper suggests a clever trick to make these simulations even more accurate. By running the simulation with different "whip strengths" and then using a mathematical technique called "extrapolation" (similar to how you might guess the true temperature by measuring it at different times of day), you can cancel out the errors. This allows them to estimate the final result with much higher precision without needing to run the simulation for an impossibly long time.

Why This Changes the Rules

This discovery has some mind-bending implications for the future of quantum computing and physics:

Pure Noise Can Compute: The paper proves that a system driven only by noise (dissipation) is powerful enough to perform any calculation a standard quantum computer can do. This means you don't necessarily need a perfect, isolated quantum system to do complex math; a noisy, open system might be enough if you know how to "whip" it right.
Freezing Time with Noise: Usually, noise makes things chaotic. But the authors show that if you apply the right kind of noise, you can actually "freeze" a system in place, stopping it from evolving at all. This is a new kind of "Zeno effect," where the environment doesn't just disturb the system, but actively cancels out its natural motion.
The Speed Limit: The paper also confirms that you can't magically speed up these noisy simulations. Just as you can't drive a car faster than the speed of light, you can't simulate these dissipative systems faster than a certain quadratic limit. This sets a hard boundary on how fast we can process information using these noisy methods.

In short, this paper flips the script on how we view quantum noise. Instead of seeing it as a destructive force that ruins our calculations, the authors show that with the right perspective, noise can be the very engine that drives quantum dynamics. It's a reminder that in the quantum world, sometimes the best way to move forward is to lean into the chaos.

Technical Summary: Hamiltonian Dynamics from Pure Dissipation

Problem Statement
The fundamental distinction between closed and open quantum systems lies in their interaction with the environment. Closed systems evolve reversibly via unitary Hamiltonian dynamics, preserving quantum information, while open systems undergo irreversible dissipation and decoherence due to coupling with an external bath, described by the Gorini–Kossakowski–Sudarshan–Lindblad (GKSL) master equation. A central theoretical question arises: Can the coherent, reversible dynamics of a closed system (governed by a non-zero Hamiltonian $H$ ) be approximately reproduced by a purely dissipative open system (governed by jump operators $\{F_i\}$ with zero explicit Hamiltonian)? While Stinespring dilation allows open systems to be modeled as closed systems by expanding the Hilbert space, this work investigates whether the reverse is possible—simulating internal Hamiltonian dynamics using only external, strictly dissipative mechanisms—without artificially redefining the system-environment partition.

Methodology
The authors propose a constructive approach to simulate Hamiltonian dynamics using a purely dissipative Lindbladian. The core construction involves defining a single-jump operator $F$ that approximates the identity minus a scaled Hamiltonian term:
$F = I - i\delta H$
Substituting this into the Lindblad equation $\dot{\rho} = F\rho F^\dagger - \frac{1}{2}\{F^\dagger F, \rho\}$ yields:
$\frac{d\rho}{dt} = -i\delta[H, \rho] + \delta^2 \left( H\rho H - \frac{1}{2}\{H^2, \rho\} \right)$
Here, the first term represents the target Hamiltonian dynamics with strength $\delta$ , while the second term represents an unavoidable dissipative error of strength $\delta^2$ . By setting the time step $\delta = \epsilon/t$ and evolving the system for a total time $T$ , the authors utilize Duhamel's Identity to bound the diamond distance between the dissipative evolution and the target unitary evolution.

The methodology is extended to:

General Jump Operators: Showing that any smooth family of purely dissipative Lindbladians mimicking Hamiltonian dynamics must take the form $F = I - \delta A + O(\delta^2)$ where the anti-Hermitian part of $A$ is $H$ .
Local Dissipation: Decomposing a global Hamiltonian $H = \sum_i H_i$ into local terms, assigning local jump operators $F_i = I - i\delta H_i$ , and demonstrating that coherent many-body dynamics can be generated from local primitives.
Autonomous QDRIFT: Adapting the QDRIFT protocol for Hamiltonian simulation to a purely dissipative setting using displaced Pauli jump operators, where randomness is supplied by the bath rather than a classical scheduler.
Extrapolation: Applying Richardson extrapolation to expectation values to exponentially improve the dependence on the error parameter $\epsilon$ for observable estimation, distinct from channel simulation.

Key Results

Theorem 1 (Upper Bound): The authors prove that a purely dissipative Lindbladian can approximate $t$ -time Hamiltonian evolution within an error $\epsilon$ (in diamond norm) using an evolution time $T = O(\|H\|_\infty^2 t^2 / \epsilon)$ . This scaling is efficient, lacking exponential overheads in time, system size, or precision.
Theorem 2 (Lower Bound & Optimality): A geometric lower bound is established showing that the $O(t^2/\epsilon)$ scaling is necessary and optimal in the worst case. By mapping the dynamics to the Bloch sphere, the authors demonstrate a trade-off between rotational speed (Hamiltonian dynamics) and radial contraction (decoherence). To maintain a uniform-in-time error $\epsilon$ , the radial decoherence must be suppressed, forcing the total evolution time to scale quadratically with $t$ . This result captures the fundamental "decoherence cost" of catching up with Hamiltonian speed.
Autonomous QDRIFT: An autonomous protocol is constructed where jump labels are sampled with probabilities proportional to the Hamiltonian coefficients, matching the standard QDRIFT scaling of $O(t^2/\epsilon)$ for channel simulation.
Extrapolation Improvement: For the specific task of estimating expectation values (rather than simulating the channel itself), the authors show that Richardson extrapolation can reduce the required evolution time to $O(t^2 \log(1/\epsilon))$ , circumventing the linear $1/\epsilon$ dependence found in channel simulation.

Implications and Significance
The paper claims several significant theoretical implications:

BQP-Completeness: Purely dissipative dynamics are shown to be BQP-complete even before reaching approximate equilibrium. This contrasts with prior results where BQP-completeness relied on encoding computation into steady states or required complex auxiliary "timer" gadgets. Here, universality is achieved via the direct simulation of unitary dynamics.
Novel Zeno-Adjacent Freezing: The work identifies a new mechanism for the Quantum Zeno effect. Unlike the traditional Zeno effect, which freezes states within a specific Zeno subspace by strong dissipation, this mechanism actively generates a time-reversal via dissipation to cancel system evolution, capable of freezing any state.
No Super-Quadratic Fast-Forwarding: The authors establish a "no super-quadratic fast-forwarding" theorem for this class of purely dissipative Lindbladians. If such dynamics could be simulated in sub-quadratic time, it would contradict known query lower bounds for sparse parity check problems.
Gauge Freedom in Simulation: The results highlight that the simulation cost of a Lindbladian on a quantum computer depends on the choice of gauge (the specific representation of $H$ and $\{F_i\}$ ). By transforming a system with $H=0$ and large jumps into a gauge with non-zero $H$ and smaller jumps, the simulation cost (proportional to norms of operators) can be drastically reduced.

Conclusion
The paper demonstrates that internal coherent Hamiltonian dynamics can be "faked" by external pure dissipation with a polynomial overhead. The construction is shown to be geometrically optimal, revealing a fundamental trade-off between rotational speed and decoherence. These findings bridge the gap between closed and open quantum dynamics, offering new perspectives on the complexity of dissipative systems, the nature of the Zeno effect, and efficient simulation strategies for open quantum systems.