Fubini-Study Speed and Short-Time Identification of Bosonic Instability
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.
Kai: We're looking at the paper: "Fubini-Study Speed and Short-Time Identification of Bosonic Instability".
Mira: Here's the gist of the paper. Quantum-geometric speed and dynamical instability are two natural rates for a driven quantum system, and their relation is unsettled even for exactly solvable dynamics Here we show that for any autonomous multimode quadratic bosonic system referred to the baremode vacuum the Fubini–Study speed vFS is the Frobenius norm of the symmetric, stretching part of the flow This identity turns a classical stability estimate into a quantum-geometric bound, λmax ≤ √2 vFS, sharp at every mode number and saturated by a resonant pure squeezer The bound is not invertible: in a detuned parametric amplifier we hold either rate fixed while varying the other The Core Identity The central finding establishes that for pure states, the infinitesimal Fubini–Study distance follows from the quantum geometric tensor and defines vFS = dsFS/dt The Anandan–Aharonov relation gives the same rate as vFS = ΔĤ/ħ For mixed states the corresponding speed is ½√FQ with FQ ≤ 4ΔĤ2/ħ2 The decomposition theorem shows that for every N and every real symmetric G, the Fubini–Study speed at the bare-mode vacuum obeys vFS2 = Tr G2 + Tr M2, which is equal to 2‖S‖F2 where S ≡ (M + MT)/2 This result depends on the symmetric part of the flow alone.
Kai: Bounding the Instability Exponent Theorem 2 provides a sharp bound relating dynamical instability to geometric speed, stating λmax ≤ √2 vFS The constant √2 is sharp for every N This is proven by Bendixson’s theorem, which bounds the real part of every eigenvalue of M by the largest eigenvalue μmax(S) of its symmetric part The identification ‖S‖F = 2vFS precedes this step, turning a classical stability ceiling into a quantum distinguishability bound.
Mira: Reciprocal Families and Separation The reciprocal families in a parametric amplifier demonstrate the separation between the two rates under controlled variation of pump amplitude and detuning Family A varies δ holding ε2 − δ2 = λ02, causing √2 vFS to grow while λ = λ0 remains fixed Family B varies δ at fixed ε, causing vFS to be constant while λ/ε falls to zero at threshold This shows that the detuning of a parametric amplifier is directly tunable and independently measurable.
Kai: Physical Interpretation and Limitations The separation is one-sided, and the surviving direction is sharp The geometric speed mixes a spectral invariant of the flow with a term depending on the state it is referred to, so converting it into a dynamical rate requires fixing that state The passive sector of the Hamiltonian is invisible to vFS yet moves the spectrum The open system is where the present treatment is thinnest, as loss enters below only as a uniform shift of the flow exponents Experimental Feasibility The experimental window in a lossy resonator shows that for Family B, vFS/2π = 35.36 MHz identically at every row The two observables are read in separated time windows The closed-system exponent equals the loss rate at the last row, where the deterministic net growth vanishes.
Title and authors: Kai: So we're talking about "Fubini-Study Speed and Short-Time Identification of Bosonic Instability." This paper by Tishina and collaborators looks at how fast a driven quantum system can become unstable, comparing it to the Fubini-Study speed. It addresses this relation, which is still kind of unsettled even for systems that are exactly solvable.
Mira: Yeah, the title itself is pretty direct. They are trying to define a clear link between two different ways we measure how fast a quantum system moves away from its starting point, one being the Fubini-Study speed and the other being the dynamical instability rate. It sounds like they're trying to set a new benchmark for understanding quantum evolution in driven systems.
Lev: From an engineering standpoint, this is interesting because it connects a purely geometric concept, the Fubini-Study speed, to something we can actually measure in terms of how fast things grow or decay in the system. That connection is what makes it relevant for building controllers or even just understanding noise in our systems.
Kai: Exactly. The main finding they present is that for any autonomous multimode quadratic bosonic system, specifically looking at the bare-mode vacuum, this Fubini-Study speed is actually the Frobenius norm of the symmetric, stretching part of the flow. This means it's a way to measure how fast the vacuum itself becomes distinguishable from itself.
Mira: And that identity turns a classical stability estimate into a quantum geometric bound. They establish that you get an upper bound on your dynamical instability rate, lambda max, which is less than or equal to the square root of two times this Fubini-Study speed. This bound is sharp for every single mode number, and they show it gets hit by a resonant pure squeezer.
Lev: The part about it being sharp at every mode number is pretty strong because that means you can't just have a loose bound that works for some modes but not others. And the saturation by a resonant pure squeezer tells us exactly what kind of dynamics creates this limit, which is super useful for designing experiments to test these limits.
Kai: They also point out something important about the invertibility of this bound. They say it's not invertible; in a detuned parametric amplifier, you can hold either rate fixed while varying the other one. That means you can control both the speed and the instability independently under different conditions for that amplifier setup.
Mira: That independence is what I find really compelling because it shows that these two rates aren't just two sides of the same coin, but they offer different ways to probe the system's behavior in distinct ways. The paper also mentions splitting the Gaussian quantum Fisher information into a sector tracking the symplectic spectrum and another one associated with correlation generating dynamics.
Title and authors: Lev: That split is key for understanding where the dynamics are really happening. The first sector vanishes on the pure-state manifold, which makes sense because that's where we're starting from, but the second sector is what captures things like how correlations build up, and it doesn't vanish for evolution generating dynamics.
Kai: So when they look at mixed states, the corresponding speed is related to the square root of two times some quantity F Q, which is bounded by four times the square of the energy uncertainty divided by two h bar squared. This shows how that same geometric idea applies when you move away from a pure state.
Mira: It connects back to what we see in other papers on quantum speed limits, but they're applying it to this specific structure of quadratic bosonic systems. The decomposition theorem they use shows that the square of the Fubini-Study speed is equal to the trace of G squared plus the trace of M squared, which equals two times the Frobenius norm squared.
Lev: That decomposition really simplifies things for analysis because it boils down to just looking at that symmetric part of the flow, M plus its transpose. That's a lot cleaner than dealing with all those cross terms in the full matrix formulation.
Kai: And that leads directly to Theorem two, which gives us the sharp bound on instability again: lambda max is less than or equal to square root of two times vFS. The constant square root of two is sharp for every number of modes, and they prove it using Bendixson's theorem to bound the real part of eigenvalues.
Mira: That use of Bendixson's theorem to bound the real part is a nice touch because it connects this geometric speed concept back to classical stability ceilings. It turns a classical stability ceiling into a quantum distinguishability bound, which is what I like about it.
Lev: But we have to be careful about that conversion. The text explicitly says that the identification, vFS equals one quarter of the Frobenius norm of M plus its transpose, converts that classical stability ceiling into this quantum distinguishability bound. That step is where we need to trust the math a lot because it's a non-trivial bridge.
Kai: So they then show how this leads to reciprocal families in a parametric amplifier. They show Family A where they vary delta while holding epsilon squared minus delta squared equal to lambda zero squared, causing the square root of two times vFS to grow, while lambda zero stays fixed.
Mira: And Family B is interesting because when you vary delta at a fixed epsilon, the Fubini-Study speed remains constant, but the instability rate lambda over epsilon drops down to zero right at some threshold. This demonstrates that detuning in a parametric amplifier is directly tunable and can be measured independently of the other rates.
Title and authors: Lev: That’s a strong result because it shows you have two distinct observables that evolve differently under these conditions. If you can control one rate while keeping the other fixed, you're essentially setting up an independent measurement of something physical.
Kai: The paper is very clear about where this treatment is weakest, though. They say the open system part is where it gets thinnest because loss enters below only as a uniform shift of the flow exponents, and for that part, they have to stick to the closed-system value of Fubini-Study speed.
Mira: That limitation on loss is expected for now because modeling genuine dissipation introduces a Lindblad generator, which drives the state off the pure-state manifold. So they have to treat loss as an external thing that shifts the whole flow structure rather than something that changes the fundamental geometric speed itself.
Lev: When you look at their experimental feasibility, they found in a lossy resonator that for Family B, vFS over pi squared is exactly thirty-five point three six mega hertz across every row. That's a concrete number derived from the model.
Kai: And they say you can read the two observables in separate time windows in that lossy system. The closed-system exponent equals the loss rate at the very last row, where it goes to zero because of deterministic net growth vanishing there.
Mira: So what this means for us is that we have a way to measure both the exponential growth rate and its amplitude separately, depending on which conditions we hold fixed in a parametric amplifier setup. It gives us two distinct features of the same trajectory.
Lev: For running this on real hardware, it implies that if you are trying to build an oscillator or some kind of driven system where instability is a key concern, you have two different ways to characterize the instability that are not necessarily coupled in a simple way.
Kai: So the core idea is using the Fubini-Study speed as a geometric measure of stretching, which then provides a sharp ceiling on dynamical instability, and that ceiling can be controlled or observed separately depending on how you tune the system parameters.
Mira: It’s an important result because it brings this quantum geometric language to help us understand fundamental limits in driven systems and where those limits might be saturated in real experiments.
Lev: We'll see how this idea translates into the next generation of error correction protocols, especially since we know these bounds are sharp for every mode number. That means we have a solid theoretical limit to work against on our side of things.
Kai: Right. So that's the gist of "Fubini-Study Speed and Short-Time Identification of Bosonic Instability." It really lays out how to use geometry to bound instability in these kinds of systems.
The paper's summary: Kai: So, to recap, they're saying that we can use something called the Fubini-Study speed to measure how fast a quantum system gets unstable in a way that links classical stability limits to quantum geometry.
Mira: Right. The core finding is this identity where the Fubini-Study speed is actually just the norm of the symmetric, stretching part of how the flow moves. That turns what sounds like a classical stability estimate into something about how distinguishable two quantum states become over time.
Lev: From an engineering standpoint, that means we're looking at a geometric measure of stretching in the evolution itself, which is useful for understanding how fast errors can grow when you're trying to keep a quantum system stable.
Kai: The paper sets up this sharp bound on instability, lambda max is at most the square root of two times this speed, and it’s sharp for every single mode number in a multimode quadratic bosonic system. It gets saturated when you have a resonant pure squeezer.
Mira: That saturation by a specific type of dynamics tells us exactly what kind of noise or driving you need to look for to hit that limit. And they show this bound isn't invertible, meaning we can control the instability rate and the geometric speed independently in things like parametric amplifiers.
Lev: That independence is interesting for error correction because it suggests you might be able to tune different aspects of the system's evolution separately, which could affect how you manage noise.
Kai: They also point out that when you look at mixed states, the corresponding speed gets related to the square root of two times some quantity F Q which is bounded by four times the square of your energy uncertainty divided by two h bar squared. That shows this geometric idea still applies even when the state isn't perfectly pure.
Mira: It connects back to how we think about quantum limits, but they are applying it specifically to these quadratic bosonic systems where you have a very structured Hamiltonian. The decomposition theorem they use is important because it boils down the full speed calculation into just looking at the symmetric part of the flow, M plus its transpose.
Lev: That simplification is necessary for running this on hardware; we need to strip away all that complexity and focus on what's actually driving the growth, which seems to be that symmetric part.
Kai: So they prove that by looking at just the symmetric part of the flow, you capture everything needed for this bound, turning a classical stability ceiling into a quantum distinguishability limit. It's a neat way to frame it.
Mira: And then they use Bendixson's theorem to show why the square root of two constant is sharp for every mode number, which ties that geometric speed back to bounding the real part of eigenvalues in a classical sense.
The paper's summary: Lev: That bridge from classical stability ceilings into quantum bounds is a crucial step; it validates the geometry by showing its connection to something we already understand about dynamical systems.
Kai: The reciprocal families they demonstrate in parametric amplifiers are really what make this paper interesting for experimentalists. Family A shows you can keep one rate fixed while the geometric speed grows, and Family B shows you can hold the geometric speed constant while the instability rate drops to zero at a threshold.
Mira: That separation is key because it means these two rates give us two different ways to probe the system's behavior under controlled variation of parameters like pump amplitude and detuning. It suggests that we can measure both features of one trajectory separately.
Lev: If you can control one rate while keeping the other fixed, it means you have a way to set up an independent measurement of something physical in your driven system. That’s a powerful setup for testing these limits on real hardware.
Kai: The paper does flag that the open system part, where loss enters, is where this treatment gets thinnest because loss just acts as a uniform shift in the flow exponents rather than changing the fundamental geometric speed itself.
Mira: That's because once you introduce true dissipation with a Lindblad generator, you move off the pure-state manifold and that changes what the Bures speed even means, so they have to treat it separately for now.
Lev: For real hardware, that means we need to be careful when we look at loss; it's not just adding a simple damping term in the same way, it fundamentally changes the nature of the evolution you're tracking.
Kai: In terms of experimental results, they found in a lossy resonator that for Family B, this geometric speed is exactly thirty-five point three six mega hertz across every row of data. That’s a concrete number from their model.
Mira: And they say you can read these two observables—the exponential growth rate and its amplitude—in separate time windows in that lossy system. It shows the geometry is encoded in both features simultaneously, just probed differently depending on your setup.
Lev: So what this means for those building quantum systems is that you have a way to characterize instability that isn't necessarily coupled in a simple way, which informs how you design feedback loops or error correction protocols.
Kai: The whole point is using the Fubini-Study speed as a geometric measure of stretching, which then gives us a sharp ceiling on instability, and we can control or observe that ceiling separately by tuning the system parameters.
Mira: It’s an important result because it brings this quantum geometric language to help us understand fundamental limits in driven systems and where those limits might actually be hit in experiments.
Lev: We'll see how this idea translates into the next generation of error correction protocols, especially since we know these bounds are sharp for every mode number, setting a solid theoretical limit on our side.
The paper's improvements: Tom: So we're looking at how the authors suggest ways to make this paper better, and what that means for future work on these driven systems.
Kai: The main suggestion is to use the Fubini-Study speed as a direct measure of instantaneous symplectic stretching in the flow, which should help us calculate that instability bound more accurately.
Mira: That makes sense because it moves the measurement closer to what we intuitively think about geometric evolution, but it relies on this idea that you can decompose the flow into symmetric and antisymmetric parts to capture all the dynamics.
Lev: From an error correction standpoint, if they get a better way to calculate that stretching part, it could mean we can predict when an error will grow fastest in a system before we even run the full simulation.
Kai: They also show how this geometric speed helps us distinguish between transient and asymptotic behavior by looking at how the antisymmetric part of the flow rotates while the symmetric part stretches. It turns that classical stability ceiling into a quantum distinguishability bound.
Mira: That distinction is important because it tells us exactly when our initial assumptions about stability might break down and when we need to look for something new in the dynamics.
Lev: If we can reliably identify those transition points, it gives us better control parameters to set for things like error correction protocols that depend on time scales.
Kai: They also point out that the open system part, where loss enters, is where they think we need more work because loss just acts as a uniform shift in the flow exponents rather than changing the core geometric speed.
Mira: So they are suggesting that for a complete picture, we should focus on how those uniform shifts interact with the pure-state evolution, even if they can't perfectly capture genuine dissipation yet.
Lev: That means any real hardware implementation will need to account for that uniform shift when mapping their measured instability rate back to the closed-system speed they calculated.
Kai: The experimental feasibility part also implies that in a lossy resonator, you can read the two observables—the growth rate and the geometric speed—in separate time windows, which is a practical way to test this separation.
Mira: It suggests that we don't necessarily need one perfect measurement to get both rates; we just need to design your experiment with two different time windows tuned specifically for each rate.
Lev: That's a practical implication for our hardware work; it means the readout system doesn't have to be perfectly synchronized across all dynamics simultaneously.
Kai: Overall, these improvements suggest that using geometry is a robust way to set these limits, even when dealing with complex, driven systems and lossy environments.
Mira: It's a solid framework for connecting the mathematical structure of the Hamiltonian directly to the observable dynamical behavior of the quantum state.
Lev: This kind of theoretical foundation will be really helpful when we start designing actual error correction architectures that need to account for these fundamental geometric limits.
Conclusion: Kai: So to wrap up, this paper on "Fubini-Study Speed and Short-Time Identification of Bosonic Instability" shows us how to use quantum geometry—specifically the speed of evolution—to set sharp limits on how fast a driven system can become unstable.
Mira: Exactly. The big implication is that we have a new way to see the connection between geometric structure and dynamical instability, which should guide our experiments in building robust quantum systems.
Lev: For us building real hardware, knowing these bounds are sharp for every mode number gives us a solid theoretical limit to work against when designing feedback loops or error correction protocols.
Kai: The key result is that we can control the instability rate and the geometric speed independently in certain setups, which opens up new ways to characterize system behavior.
Mira: That separation between the two rates is what makes this paper compelling; it means we get two different pieces of information from one trajectory, depending on how we tune our amplifier or resonator.
Lev: I just think that ability to decouple those variables is crucial when you're trying to manage the overhead of error correction in a noisy environment.
Kai: It’s a neat result because it connects the pure mathematics of quantum geometry right into the physical world of how fast things grow or decay in our circuits.
Mira: We need to keep looking at how this geometric speed relates to loss, because that open system treatment is where they flag the biggest caveat right now.
Lev: Yeah, if we can nail down that uniform shift from loss, then we'll have a much clearer picture of what those numbers actually mean in a real experimental setup.
Lomonosov Moscow State University · Moscow Institute of Physics and Technology
cond-mat.mes-hall, quant-ph
Submitted: 2026-08-11
Updated: 2026-11-11
Comments: 25 pages 3 fig 2 tables
Code: https://github.com/alexandermtishin-art/IntrinsicGeometric-Energy-of-Driven-Quantum-Evolution
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
Importance score: 89/100
The gist: Quantum-geometric speed and dynamical instability are two natural rates for a driven quantum system, and their relation is unsettled even for exactly solvable dynamics Here we show that for any…
Key concepts
- Fubini–Study speed (vFS)
- This is a measure of how quickly the quantum state's distance changes over time, derived from the quantum geometric tensor. It is equivalent to the rate at which the Hamiltonian evolves a pure state, providing a geometric bound on system dynamics.
- Dynamical Instability ($\lambda$)
- This represents the exponential growth rate of perturbations in a driven quantum system. It quantifies how quickly small deviations from an equilibrium state will grow over time, often related to classical stability limits.
- Symmetric Part of the Flow (G)
- The flow describes how quantum states evolve. The symmetric part of this flow is crucial because it represents the stretching or growth component of the dynamics. The Fubini-Study speed is directly proportional to the Frobenius norm of this symmetric, stretching part.
- Resonant Pure Squeezer
- This specific state configuration saturates the theoretical bound ($\lambda_{max} = \sqrt{2} vFS$). It represents a system operating at a critical point where quantum geometric speed reaches its maximum possible value for that mode number.
Terminology
Summary
Quantum-geometric speed and dynamical instability are two natural rates for a driven quantum system, and their relation is unsettled even for exactly solvable dynamics Here we show that for any autonomous multimode quadratic bosonic system referred to the baremode vacuum the Fubini–Study speed vFS is the Frobenius norm of the symmetric, stretching part of the flow This identity turns a classical stability estimate into a quantum-geometric bound, λmax ≤ √2 vFS, sharp at every mode number and saturated by a resonant pure squeezer The bound is not invertible: in a detuned parametric amplifier we hold either rate fixed while varying the other
The Core Identity
The central finding establishes that for pure states, the infinitesimal Fubini–Study distance follows from the quantum geometric tensor and defines vFS = dsFS/dt The Anandan–Aharonov relation gives the same rate as vFS = ΔĤ/ħ For mixed states the corresponding speed is ½√FQ with FQ ≤ 4ΔĤ2/ħ2 The decomposition theorem shows that for every N and every real symmetric G, the Fubini–Study speed at the bare-mode vacuum obeys vFS2 = Tr G2 + Tr M2, which is equal to 2‖S‖F2 where S ≡ (M + MT)/2 This result depends on the symmetric part of the flow alone.
Bounding the Instability Exponent
Theorem 2 provides a sharp bound relating dynamical instability to geometric speed, stating λmax ≤ √2 vFS The constant √2 is sharp for every N This is proven by Bendixson’s theorem, which bounds the real part of every eigenvalue of M by the largest eigenvalue μmax(S) of its symmetric part The identification ‖S‖F = 2vFS precedes this step, turning a classical stability ceiling into a quantum distinguishability bound.
Reciprocal Families and Separation
The reciprocal families in a parametric amplifier demonstrate the separation between the two rates under controlled variation of pump amplitude and detuning Family A varies δ holding ε2 − δ2 = λ02, causing √2 vFS to grow while λ = λ0 remains fixed Family B varies δ at fixed ε, causing vFS to be constant while λ/ε falls to zero at threshold This shows that the detuning of a parametric amplifier is directly tunable and independently measurable.
Physical Interpretation and Limitations
The separation is one-sided, and the surviving direction is sharp The geometric speed mixes a spectral invariant of the flow with a term depending on the state it is referred to, so converting it into a dynamical rate requires fixing that state The passive sector of the Hamiltonian is invisible to vFS yet moves the spectrum The open system is where the present treatment is thinnest, as loss enters below only as a uniform shift of the flow exponents
Experimental Feasibility
The experimental window in a lossy resonator shows that for Family B, vFS/2π = 35.36 MHz identically at every row The two observables are read in separated time windows The closed-system exponent equals the loss rate at the last row, where the deterministic net growth vanishes. p. 8 In a quartic extension of the twisting-and-turning
Hamiltonian, two models tuned to nearly equal Lyapunov exponents differ in metrological
gain, the short-time growth of the anti-squeezed variance being set by a rate built from the
two local curvatures rather than by the exponent itself [17]. In the notation used here that rate
is the Frobenius norm of the symmetric part of the same linearized flow, so it equals √2 vFS,
and their two expressions give in one step that the difference of its square from the squared
exponent is the square of the curvature sum, hence non-negative, which is Eq.Where the paper says it
Competing Interests
The author declares no competing interests The Python implementations reproducing every number in this article and in the Supplemental Material are available at https://github.com/alexandermtishin-art/IntrinsicGeometric-Energy-of-Driven-Quantum-Evolution During the preparation of this work the author used generative AI tools to assist with language editing, notation consistency, reference cross-checking, and Python code development The author reviewed and edited the output and takes full responsibility for the content of the publication.
References
[1] J. P. Provost and G. Vallée, Riemannian structure on manifolds of quantum states, Commun. p. 0 —The spectrum of a Hamiltonian says which energies are available; it does not
say how fast the state moves among them. Projective Hilbert space supplies the missing rate:
the Fubini–Study line element, unique up to scale among unitarily invariant Riemannian
metrics on the manifold of quantum states [1], measures how fast a state becomes
distinguishable from its own recent past.Where the paper says it
[2] J. Anandan and Y. p. 2 —For pure states the infinitesimal Fubini–Study distance follows from the quantum
geometric tensor [1] and defines vFS = dsFS/dt, which is gauge invariant, independent of the
choice of quadrature basis in the sense made precise below, and satisfies vFS2dt
2 = 1 − F for the
fidelity F of an infinitesimal step. The Anandan–Aharonov relation gives the same rate as vFS =
ΔĤ/ℏ [2]; for mixed states the corresponding speed is ½√FQ with FQ ≤ 4ΔĤ2/ℏ2 [3,4], and
closed forms for FQ on Gaussian states are known [18–20]. —For pure states the infinitesimal Fubini–Study distance follows from the quantum
geometric tensor [1] and defines vFS = dsFS/dt, which is gauge invariant, independent of the
choice of quadrature basis in the sense made precise below, and satisfies vFS2dt
2 = 1 − F for the
fidelity F of an infinitesimal step. The Anandan–Aharonov relation gives the same rate as vFS =
ΔĤ/ℏ [2]; for mixed states the corresponding speed is ½√FQ with FQ ≤ 4ΔĤ2/ℏ2 [3,4], and
closed forms for FQ on Gaussian states are known [18–20].Where the paper says it
Where the paper says it
[3] S. L. Braunstein and C. M. Caves, Statistical distance and the geometry of quantum states, Phys. p. 2 The step taken here is that identification, vFS = ¼‖M +
MT‖F, which converts a classical stability ceiling into a quantum distinguishability bound and
makes the two rates independently controllable in one parametric amplifier. —For pure states the infinitesimal Fubini–Study distance follows from the quantum
geometric tensor [1] and defines vFS = dsFS/dt, which is gauge invariant, independent of the
choice of quadrature basis in the sense made precise below, and satisfies vFS2dt
2 = 1 − F for the
fidelity F of an infinitesimal step.Where the paper says it
[4] M. M. Taddei, B. M. Escher, L. Davidovich and R. L. de Matos Filho, Quantum speed limit for physical processes, Phys. p. 0 Projective Hilbert space supplies the missing rate:
the Fubini–Study line element, unique up to scale among unitarily invariant Riemannian
metrics on the manifold of quantum states [1], measures how fast a state becomes
distinguishable from its own recent past. For pure states that rate is the energy uncertainty
divided by ℏ [2]; for mixed states it is set by the quantum Fisher information [3,4], and it
controls the Mandelstam–Tamm speed limit [5] and its descendants [6].Where the paper says it
[5] L. Mandelstam and I. p. 0 Projective Hilbert space supplies the missing rate:
the Fubini–Study line element, unique up to scale among unitarily invariant Riemannian
metrics on the manifold of quantum states [1], measures how fast a state becomes
distinguishable from its own recent past. For pure states that rate is the energy uncertainty
divided by ℏ [2]; for mixed states it is set by the quantum Fisher information [3,4], and it
controls the Mandelstam–Tamm speed limit [5] and its descendants [6]. Projective Hilbert space supplies the missing rate:
the Fubini–Study line element, unique up to scale among unitarily invariant Riemannian
metrics on the manifold of quantum states [1], measures how fast a state becomes
distinguishable from its own recent past. For pure states that rate is the energy uncertainty
divided by ℏ [2]; for mixed states it is set by the quantum Fisher information [3,4], and it
controls the Mandelstam–Tamm speed limit [5] and its descendants [6].Where the paper says it
Where the paper says it
[6] S. Deffner and S. Campbell, Quantum speed limits: from Heisenberg’s uncertainty principle to optimal quantum control, J. p. 8 What Theorem 2 adds is the identification of the rate with a quantum distinguishability speed,
the extension to any mode number with the constant √2 shown to be sharp, and the
characterization of the passive sector that makes the bound non-invertible. Limits of instantaneous spectral diagnostics.Where the paper says it
[7] P. M. Poggi, S. Campbell and S. Deffner, Diverging quantum speed limits: a herald of classicality, PRX Quantum 2, 040349 (2021). p. 8 What Theorem 2 adds is the identification of the rate with a quantum distinguishability speed,
the extension to any mode number with the constant √2 shown to be sharp, and the
characterization of the passive sector that makes the bound non-invertible. Limits of instantaneous spectral diagnostics.Where the paper says it
[8] K. Chatterjee, T. Pandit, V. Singh, P. Chattopadhyay and U. L. Andersen, Even-odd splitting of the Gaussian quantum Fisher information: from symplectic geometry to metrology, arXiv:2601 06513 (2026). p. 3 (2) it contributes nothing to vFS while entering the spectrum of M. That a purely passive generator produces no geometric motion of the vacuum reference used
here follows from the even–odd decomposition of the Gaussian quantum Fisher information
[8] and is not claimed here as new; Eq.Where the paper says it
[9] S. Das, Geometry- and inertia-limited chaotic growth in classical many-body systems, arXiv:2602 21149 (2026). p. 1 Closest of all, in a collective-spin sensing
context the short-time growth of the anti-squeezed variance near a hyperbolic point has been
shown to be governed by a rate distinct from the Lyapunov exponent, so that two systems
tuned to the same exponent amplify differently [17]. We are not aware of a treatment that identifies that classical stability object with the Fubini–
Study speed of the quantum vacuum, or that converts the non-normality diagnosis into aWhere the paper says it
[10] K. M. Fonseca Romero, J. E. Parreira, L. A. M. Souza, M. C. Nemes and W. F. Wreszinski, Analytical relation between entropy production and quantum Lyapunov exponents for Gaussian bipartite systems, J. p. 1 Independently, upper
bounds on Hamiltonian chaotic growth have been derived from instantaneous stability
constraints set by the largest eigenvalue of the symmetric part of the stability matrix, with no
statistical assumptions and no long-time averaging [9]. Information-theoretic quantities have
been connected to instability before, in unstable Gaussian environments where coherence loss
is determined by the upper quantum Lyapunov exponent [10], and a quantum speed limit for
the out-of-time-ordered correlator gives an upper bound on quantum Lyapunov exponents in
terms of environmental two-point correlation functions [11].Where the paper says it
[11] D. Tripathy, J. Thingna and S. Deffner, Quantum speed limit for the OTOC from an open systems perspective, arXiv:2509 13519 (2025). p. 1 Closest of all, in a collective-spin sensing
context the short-time growth of the anti-squeezed variance near a hyperbolic point has been
shown to be governed by a rate distinct from the Lyapunov exponent, so that two systems
tuned to the same exponent amplify differently [17]. We are not aware of a treatment that identifies that classical stability object with the Fubini–
Study speed of the quantum vacuum, or that converts the non-normality diagnosis into aWhere the paper says it
[12] V. P. Flynn, E. Cobanera and L. Viola, Deconstructing effective non-Hermitian dynamics in quadratic bosonic Hamiltonians, New J. p. 1 Information-theoretic quantities have
been connected to instability before, in unstable Gaussian environments where coherence loss
is determined by the upper quantum Lyapunov exponent [10], and a quantum speed limit for
the out-of-time-ordered correlator gives an upper bound on quantum Lyapunov exponents in
terms of environmental two-point correlation functions [11]. The non-normality that
separates transient from asymptotic behaviour has been analysed for closed quadratic bosonic
Hamiltonians [12] and, by pseudospectral methods, for their Markovian extensions, where it
classifies the metastable regimes that follow [13–16].Where the paper says it
[13] V. P. Flynn, E. Cobanera and L. Viola, Topology by dissipation: Majorana bosons in metastable quadratic Markovian dynamics, Phys. p. 1 Information-theoretic quantities have
been connected to instability before, in unstable Gaussian environments where coherence loss
is determined by the upper quantum Lyapunov exponent [10], and a quantum speed limit for
the out-of-time-ordered correlator gives an upper bound on quantum Lyapunov exponents in
terms of environmental two-point correlation functions [11]. The non-normality that
separates transient from asymptotic behaviour has been analysed for closed quadratic bosonic
Hamiltonians [12] and, by pseudospectral methods, for their Markovian extensions, where it
classifies the metastable regimes that follow [13–16].Where the paper says it
[14] V. P. Flynn, E. Cobanera and L. Viola, Topological zero modes and edge symmetries of metastable Markovian bosonic systems, Phys. p. 1 The non-normality that
separates transient from asymptotic behaviour has been analysed for closed quadratic bosonic
Hamiltonians [12] and, by pseudospectral methods, for their Markovian extensions, where it
classifies the metastable regimes that follow [13–16]. Closest of all, in a collective-spin sensing
context the short-time growth of the anti-squeezed variance near a hyperbolic point has been
shown to be governed by a rate distinct from the Lyapunov exponent, so that two systems
tuned to the same exponent amplify differently [17].Where the paper says it
[15] M. Ughrelidze, V. P. Flynn, E. Cobanera and L. Viola, Interplay of finite- and infinite-size stability in quadratic bosonic Lindbladians, Phys. p. 0 Here we
show that for any autonomous multimode quadratic bosonic system referred to the baremode vacuum the Fubini–Study speed vFS is the Frobenius norm of the symmetric, stretching
part of the flow: the rate at which the vacuum becomes distinguishable from itself measures
instantaneous symplectic stretching. That identity turns a classical stability estimate into a
quantum-geometric bound, λmax ≤ √2 vFS, sharp at every mode number and saturated by a
resonant pure squeezer.Where the paper says it
[16] V. P. Flynn, Effective Non-Hermiticity and Topology in Markovian Quadratic Bosonic Dynamics, Springer Theses (Springer, Cham, 2024). p. 1 Information-theoretic quantities have
been connected to instability before, in unstable Gaussian environments where coherence loss
is determined by the upper quantum Lyapunov exponent [10], and a quantum speed limit for
the out-of-time-ordered correlator gives an upper bound on quantum Lyapunov exponents in
terms of environmental two-point correlation functions [11]. The non-normality that
separates transient from asymptotic behaviour has been analysed for closed quadratic bosonic
Hamiltonians [12] and, by pseudospectral methods, for their Markovian extensions, where it
classifies the metastable regimes that follow [13–16].Where the paper says it
[17] B. Vijaywargia, J. Chávez-Carlos, F. Pérez-Bernal and L. F. Santos, Instability-enhanced quantum sensing with tunable multibody interactions, arXiv:2604 05047 (2026). https://doi.org/10.48550/arXiv 2604 05
[18] O. Pinel, P. Jian, N. Treps, C. Fabre and D. Braun, Quantum parameter estimation using general single-mode Gaussian states, Phys. p. 9 The evolution treated in Theorem 1 is unitary, so loss enters
only through the flow exponents, where it is a uniform shift, and the fidelity decay quoted in
the Appendix is restricted to κt ≪ 1. And Theorem 1 is stated at the bare-mode vacuum: the
same variance formula gives a closed expression for any Gaussian reference, but not the same
one, since for general Σ the first term is ½ Tr(GΣGΣ) rather than ⅛ Tr G2, and vFS ceases to be a
function of ‖S‖F alone.Where the paper says it
[19] Z. Jiang, Quantum Fisher information for states in exponential form, Phys. p. 0 Projective Hilbert space supplies the missing rate:
the Fubini–Study line element, unique up to scale among unitarily invariant Riemannian
metrics on the manifold of quantum states [1], measures how fast a state becomes
distinguishable from its own recent past. For pure states that rate is the energy uncertainty
divided by ℏ [2]; for mixed states it is set by the quantum Fisher information [3,4], and it
controls the Mandelstam–Tamm speed limit [5] and its descendants [6].Where the paper says it
[20] J. Liu, H. Yuan, X.-M. Lu and X. Wang, Quantum Fisher information matrix and multiparameter estimation, J. p. 1 The Gaussian quantum Fisher information has been split
into a sector tracking the symplectic spectrum and a sector associated with correlationgenerating dynamics, the first vanishing on the pure-state manifold and the second vanishing
for evolutions generated entirely by passive Gaussian unitaries [8]. Independently, upper
bounds on Hamiltonian chaotic growth have been derived from instantaneous stability
constraints set by the largest eigenvalue of the symmetric part of the stability matrix, with no
statistical assumptions and no long-time averaging [9].Where the paper says it
[21] J. Williamson, On the algebraic problem concerning the normal forms of linear dynamical systems, Am. p. 0 1
Quantum-Geometric Bound on Dynamical Instability in Bosonic Systems
A. Tishina,b*
a Lomonosov Moscow State University, Physics Department, 119991, Leninskie gory 1,
Moscow, Russia
b Moscow Institute of Physics and Technology, 141701, Institutskiy per.Where the paper says it
[22] A. Serafini, Quantum Continuous Variables: A Primer of Theoretical Methods (CRC Press, Boca Raton, 2017). p. 1 Information-theoretic quantities have
been connected to instability before, in unstable Gaussian environments where coherence loss
is determined by the upper quantum Lyapunov exponent [10], and a quantum speed limit for
the out-of-time-ordered correlator gives an upper bound on quantum Lyapunov exponents in
terms of environmental two-point correlation functions [11]. The non-normality that
separates transient from asymptotic behaviour has been analysed for closed quadratic bosonic
Hamiltonians [12] and, by pseudospectral methods, for their Markovian extensions, where it
classifies the metastable regimes that follow [13–16].Where the paper says it
[23] S. Tanaka, G. Ordonez, M. Kubota, H. Nakano, B. Mera and K. Kanki, Operator-frame geometry of non-compact quantum systems with frame-vacuum phase transitions, arXiv:2607 06994 (2026). p. 1 Closest of all, in a collective-spin sensing
context the short-time growth of the anti-squeezed variance near a hyperbolic point has been
shown to be governed by a rate distinct from the Lyapunov exponent, so that two systems
tuned to the same exponent amplify differently [17]. We are not aware of a treatment that identifies that classical stability object with the Fubini–
Study speed of the quantum vacuum, or that converts the non-normality diagnosis into aWhere the paper says it
[24] I. Tesfaye and A. Eckardt, Quantum geometry of bosonic Bogoliubov quasiparticles, Phys. p. 3 Second, the object used here is not the symplectic quantum geometric tensor, which
supplies a metric on the space of bosonic Bogoliubov modes [24]: vFS is the speed of one fixed
state transported by the flow, so the two constructions are defined on different spaces and are
not interchangeable. The decomposition theorem.Where the paper says it
[25] See Supplemental Material [https://github.com/alexandermtishin-art/IntrinsicGeometric-Energy-of-Driven-Quantum-Evolution] for the Gaussian variance formula and its validation, the numerical verification of Theorems 1 and 2 and of the isospectral construction, the chirped mode comparison, the parametric-resonance counterexample and the closed forms for a driven thermal state, which includes Refs. p. 0 For pure states that rate is the energy uncertainty
divided by ℏ [2]; for mixed states it is set by the quantum Fisher information [3,4], and it
controls the Mandelstam–Tamm speed limit [5] and its descendants [6]. Quadratic bosonic dynamics — parametric amplifiers, chirped modes, driven condensates,
linearized collective spins — carries a second rate in the same units, the instability exponentWhere the paper says it
[26] M. H. Muñoz-Arias, I. H. Deutsch and P. M. Poggi, Phase-space geometry and optimal state preparation in quantum metrology with collective spins, PRX Quantum 4, 020314 (2023). p. 8 —The separation appears in a setting not constructed for the
purpose. For the twisting-and-turning collective-spin Hamiltonian analyzed in the context of
optimal metrological state preparation [26], the effective quadratic generator obtained by
linearizing about the unstable saddle point takes the form of Eq.Where the paper says it
[27] W. G. Hoover and C. G. Hoover, Instantaneous pairing of Lyapunov exponents in chaotic Hamiltonian dynamics and the 2017 Ian Snook Prizes, arXiv:1703 00470 (2017). p. 1 Independently, upper
bounds on Hamiltonian chaotic growth have been derived from instantaneous stability
constraints set by the largest eigenvalue of the symmetric part of the stability matrix, with no
statistical assumptions and no long-time averaging [9]. Information-theoretic quantities have
been connected to instability before, in unstable Gaussian environments where coherence loss
is determined by the upper quantum Lyapunov exponent [10], and a quantum speed limit for
the out-of-time-ordered correlator gives an upper bound on quantum Lyapunov exponents in
terms of environmental two-point correlation functions [11].Where the paper says it
[28] S. Boutin, D. M. Toyli, A. V. Venkatramani, A. W. Eddins, I Siddiqi and A Blais, Effect of higher-order nonlinearities on amplification and squeezing in Josephson parametric amplifiers, Phys. p. 0 For pure states that rate is the energy uncertainty
divided by ℏ [2]; for mixed states it is set by the quantum Fisher information [3,4], and it
controls the Mandelstam–Tamm speed limit [5] and its descendants [6]. Quadratic bosonic dynamics — parametric amplifiers, chirped modes, driven condensates,
linearized collective spins — carries a second rate in the same units, the instability exponent Two group actions must be kept apart. A symplectic change of quadratures ẑ → V−1ẑ, applied
simultaneously to generator and state, sends G → VTGV and M → V−1MV: the generator
transforms by congruence, the flow by similarity, and vFS is unchanged.Where the paper says it
Where the paper says it
[29] M. Malnou, D. A. Palken, L. R. Vale, G. C Hilton and K W Lehnert, Optimal operation of a Josephson parametric amplifier for vacuum squeezing, Phys. p. 5 Reciprocal families in a parametric amplifier. —The realizable single-mode form is a
parametrically pumped mode in the frame rotating at half the pump frequency,
Ĥ = δ(x̂2 + p̂2)/2 + (ε/2)(x̂2 − p̂2), (4)
an ordinary Josephson parametric amplifier with detuning δ and pump amplitude ε. Two group actions must be kept apart. A symplectic change of quadratures ẑ → V−1ẑ, applied
simultaneously to generator and state, sends G → VTGV and M → V−1MV: the generator
transforms by congruence, the flow by similarity, and vFS is unchanged.Where the paper says it
Where the paper says it
[30] A. M. Tishin, Quantum geometric origin of non-adiabatic instability in driven bosonic systems, arXiv:2605 22796 (2026). p. 0 9, Dolgoprudny,
Moscow Region, Russia
*tishin@amtc. org
Abstract
Quantum-geometric speed and dynamical instability are two natural rates for a driven
quantum system, and their relation is unsettled even for exactly solvable dynamics.Where the paper says it
Appendix
Including single-photon loss, the quadrature flow acquires a uniform damping term, Mopen = M − (κ/2)·1, so the exponents become −κ/2 ± λ This treats loss at the level of the flow, where it is exact; it does not carry over to the speed, since a Lindblad generator drives the state off the pure-state manifold and the relevant quantity becomes the Bures speed Everything below is therefore stated for the exponents, and the fidelity decay is quoted only in the window κt ≪ 1 where the closed-system value of vFS still applies The quadratic fidelity decay must be taken at κt ≪ 1
The author declares no competing interests Data Availability Statement The Python implementations reproducing every number in this article and in the Supplemental Material are available at https://github.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI systems can calculate a sharp bound on dynamical instability for autonomous multimode quadratic bosonic systems by using vFS as a measure of instantaneous symplectic stretching, as shown by
the Fubini–Study speed vFS is the Frobenius norm of the symmetric, stretching part of the flow.
-
The improved system can predict exactly how two rates separate in a detuned parametric amplifier by utilizing the relationship derived from Theorem 2: "The separation is between measurable trajectories. Evolving the bare-mode vacuum under Eq. (4) gives the mean occupation exactly, which assigns the two rates different roles: 'the exponent sets the rate of the exponential, and 'the geometric speed sets its amplitude.'"
-
The system can distinguish between transient and asymptotic behavior in closed quadratic bosonic Hamiltonians by identifying that
the antisymmetric part of the flow rotates, while the symmetric part stretches,
thereby converting a classical stability ceiling into aquantum distinguishability bound.
Abstract
We ask what short-time quantum distinguishability reveals about instability in quadratic bosonic systems. For a fixed autonomous quadratic Hamiltonian, the infimum of the Fubini-Study speed over centered pure Gaussian inputs equals the Euclidean norm of the positive real parts of the flow eigenvalues divided by the square root of 2. The infimum is attained exactly when the flow generator is diagonalizable. We also derive the minimum speed under a finite squeezing budget for a detuned parametric amplifier. For a lossy single-mode amplifier with a common phase-insensitive Markovian bath, two return-fidelity traces, from the bare vacuum and a calibrated squeezed vacuum, determine the growth margin, damping, and diffusion from their initial slopes and curvatures. Finite-window simulations illustrate how the traces distinguish cases on opposite sides of the instability threshold.
Sources
- Even Odd Splitting of the Gaussian Quantum Fisher Information: From Symplectic Geometry to Metrology
- Geometry- and inertia-limited chaotic growth in classical many-body systems
- An analytical relation between entropy production and quantum Lyapunov exponents for Gaussian bipartite systems
- Quantum speed limit for the OTOC from an open systems perspective
- Instability-enhanced quantum sensing with tunable multibody interactions
- Operator-frame geometry of non-compact quantum systems with frame-vacuum phase transitions
- Instantaneous Pairing of Lyapunov Exponents in Chaotic Hamiltonian Dynamics and the 2017 Ian Snook Prize
- Identifiability and endpoint reconstruction of time-dependent Jacobi Hamiltonians from projective trajectory jets
- Geometric Instability and Self-Limitation in Driven Quantum Systems
Related papers
- Coexistence of polariton bound states in the continuum and 2D radiative excitons
- Hidden Quantum-Geometric Connection between Injection and Shift Currents
- Specific absorption rate of uniaxial single-domain nanomagnets: stochastic spin dynamics versus linear response theory
- The ideal limit of rhombohedral graphene: Dynamical skyrmion lattices and their excitations
- Valley polarization of chiral excitonic bound states induced by band geometry
- On-Device Control of Electronic Friction