Sharp Entropy Cost of Coherent Leakage Across a Block Decomposition
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: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We're looking at the paper: "Sharp Entropy Cost of Coherent Leakage Across a Block Decomposition".
Title and authors: Kai: So we're looking at this paper now, "Sharp Entropy Cost of Coherent Leakage Across a Block Decomposition," and the authors are talking about how much information you lose when you try to remove that cross-block coherence using block dephasing.
Mira: Right, it sounds very technical, but basically they're trying to put a precise number on the cost of removing those correlations between different parts of a quantum system. It's about quantifying the loss in information when you apply this block dephasing map, which they call pi rho:= P rho P plus Q rho Q.
Lev: From an engineering standpoint, what I want to know is if this cost calculation actually helps us design better error correction codes. Is this a bound on how much noise we can tolerate before the system becomes useless?
Kai: Well, they establish a sharp entropy cost theorem, which is an exact one-channel variational theorem, and they get this logarithmic boundary law that tells us exactly how that cost scales with the leakage mass. It's like getting a very precise measure of the "price" of dephasing.
Mira: That logarithmic boundary law is key because it shows that as the leakage mass gets really small, say tending toward zero, the lower bound on this entropy cost actually grows logarithmically with that tiny leakage mass, which is a pretty interesting divergence.
Lev: So, if we think about running this on real hardware, does this sharp bound tell us anything practical about how much coherence we can expect to keep in a noisy environment?
Kai: They prove a midpoint Bogoliubov–Kubo–Mori coercivity estimate, which gives us an operator kernel lower bound for that entropy cost D(rho pi rho) greater than or equal to the term Tr of B star Omega minus one A, C B when the diagonal blocks are strictly positive.
Mira: That lower bound proof uses a block-sign involution, U equals P minus Q, to get a sign symmetric midpoint lower bound for the Bogoliubov–Kubo–Mori Hessian. The crucial part they establish there is that at fixed protected sector floor, leakage mass, and quadratic coherence, the minimum block-coherence entropy can be achieved by only one active two-level block.
Lev: That sounds like a simplification for hardware design. So even if we have a big complex system, the worst case for losing coherence is always concentrated in just one effective qubit structure?
Kai: Exactly. They call this the one-channel principle, meaning that for that constrained problem, the minimum is realized by a qubit-like two-level structure regardless of how many dimensions are involved. It suggests that spreading that coherence out across more channels doesn't actually lower the entropy cost in this specific scenario.
Title and authors: Mira: Then they move on to solving the constrained extremal problem, and they find that the infimum of this block-coherence entropy equals a function phi of three things: protected sector floor, leakage mass, and quadratic coherence.
Lev: What are those other things? Because if we want to use this for real hardware simulation, we need to know exactly what parameters we're controlling.
Kai: The formula they derive for the optimal floor is a star equals one minus the leakage mass minus the inverse of dP minus one times a0. That means that parameter a star is determined by those three variables.
Mira: And then they prove this minimum state can be attained by a single active two-level block plus incoherent spectators, which is what they call the one-channel extremizer result. It shows the structure of the optimal state under these constraints.
Lev: So, from my perspective on error correction overhead, it seems like we can simplify our optimization landscape to just finding that single best two-level block structure rather than searching through a high-dimensional space of blocks?
Kai: That's what Theorem IV.eight states, giving the exact version of this statement for finite dimensions <ref:2605.11024#pg2,the exact version of this statement>. It’s a very concrete result about the shape of the optimal state when you're trying to minimize that entropy cost across a block decomposition.
Mira: And they also present a corollary that gives us that logarithmic boundary barrier, which is D(rho pi rho) greater than or equal to B star F log a0 epsilon Q when the floor A is large enough and the leakage mass epsilon Q is not too small.
Lev: That divergence as the leakage mass goes to zero, that logarithmic growth with respect to epsilon Q, is something I can use to set realistic limits on how much noise we can handle in a noisy channel before the cost skyrockets.
Kai: And they point out that this divergence happens along any family of states where the cross-block coherence term B is not zero when epsilon Q goes to zero, which is important because it shows the lower bound per unit quadratic coherence diverges.
Mira: The paper also mentions a Coherence Compression Principle, which essentially says that any state with multiple coherent blocks can be mapped by a compression operation to one with only a single active block without increasing the entropy. This is the structural reason why we can collapse this high-dimensional problem down to a scalar two by two problem.
Lev: That compression principle would be incredibly helpful for building simulators or even simplifying the analysis of complex, large systems in error correction where we have too many degrees of freedom to track simultaneously?
Title and authors: Kai: It is. And they extend the midpoint BKM mechanism into finite-dimensional trace preserving Z two graded automorphisms and also under bounded strict positivity assumptions to finite trace corners of semifinite von Neumann algebras.
Mira: Those extensions show that the core mechanism holds up even when we move into those more complex mathematical settings, which makes it relevant beyond just simple finite-dimensional models. It shows the robustness of this one-channel principle in graded and corner settings.
Lev: For real hardware, that extension means we can start thinking about how this applies to systems that might have some kind of grading or non-standard structures in their underlying algebra, not just the simplest qubit setups.
Kai: So, to wrap up the main points of "Sharp Entropy Cost of Coherent Leakage Across a Block Decomposition," they give us a sharp entropy cost theorem, the BKM midpoint coercivity and that logarithmic boundary law related to leakage mass.
Mira: The main implication is the one-channel principle: the minimum block-coherence entropy under those constraints is always achieved by just one active two-level block.
Lev: And for me, it means we can use this to argue that when designing error correction for these systems, we only really need to focus on optimizing that single best qubit structure, and not worrying about distributing the coherence across many blocks unnecessarily.
Kai: Yeah, and they also show a simulation routine for an idealized uniform block dephasing orbit where the rate of entropy decrease is bounded below by two times G e to negative two G t Tr of B star Omega minus one A C B when the floor is above a zero P and leakage mass is less than or equal to a zero over a zero.
Mira: And they flag that their method doesn't cover non-uniform dephasing rates, or full modified logarithmic Sobolev inequalities, which are important for a more complete picture of how entropy dissipates under varying noise conditions.
Lev: So the limitation is that it only works well for uniform dephasing and these specific positivity assumptions; we can't directly apply it to systems where the noise isn't spread out evenly across blocks.
Kai: So, to summarize, the paper "Sharp Entropy Cost of Coherent Leakage Across a Block Decomposition" gives us a sharp entropy cost theorem, the BKM midpoint coercivity and that logarithmic boundary law related to leakage mass.
Lev: So the limitation is that it only works well for uniform dephasing and these specific positivity assumptions; we can't directly apply it to systems where the noise isn't spread out evenly across blocks.
The paper's summary: Kai: So, to recap this paper, they're not just looking at noise anymore; they’re quantifying exactly how much information you lose when you try to separate coherent parts of a quantum system using a block decomposition map.
Mira: Right, they call that loss the block-coherence entropy. It’s the informational distance between your actual state and its dephased shadow, and it finds this sharp lower bound for it.
Lev: What's interesting is how they derived this bound using something called a midpoint Bogoliubov–Kubo–Mori estimate, which gives them a way to get that operator kernel lower bound.
Kai: Exactly. And the main result is this logarithmic boundary law, which shows that as the leakage mass gets really tiny, say approaching zero, the cost of removing that coherence grows like a logarithm with respect to that tiny leakage mass.
Mira: That divergence is pretty telling because it suggests that you can never completely eliminate this cost just by making your blocks smaller; there's always this residual logarithmic penalty as you get closer to perfect block-diagonal states.
Lev: From an error correction standpoint, that logarithmic growth rate tells us the fundamental difficulty of fighting cross-block noise when the noise is spread out uniformly.
Kai: And they have this Coherence Compression Principle, which basically says any state with many coherent blocks can be compressed into one single active two-level block without increasing the entropy.
Mira: That compression principle is what makes the whole thing manageable; it’s the structural reason why you don't need to track every single block independently when you are trying to minimize that cost.
Lev: So, if we apply that principle, we can solve the high-dimensional problem by just focusing on finding that single best two-level structure and ignoring the rest of the complexity.
Kai: And they even extended this mechanism into systems with finite dimensional trace preserving Z two graded automorphisms and corners of semifinite von Neumann algebras, which shows it’s not just a simple toy model.
Mira: That extension proves this one-channel principle is robust, meaning it holds up even in more complex mathematical settings beyond the simplest qubit setups.
Lev: It means we can start thinking about applying these cost calculations to systems with grading or more intricate structures in their underlying algebras for real hardware design.
The paper's improvements: Tom: So, we're looking at how the authors suggest ways to make this block coherence cost analysis even better by fixing some of the assumptions they had to get those results.
Kai: They point out that their current work relies on uniform dephasing rates, and they flag that a more complete picture would need to include non-uniform dephasing rates.
Mira: That makes sense because in real quantum hardware, the noise isn't always spread out evenly across all the blocks you’re trying to protect.
Lev: So what does this mean for the simulation side? If we can handle non-uniform noise, it means our error correction overhead calculations won't be as optimistic as they are right now.
Kai: And they suggest using full modified logarithmic Sobolev inequalities, which would give us a more complete way to understand how entropy actually dissipates under varying noise conditions.
Mira: That goes deeper than just the lower bound they proved; it tackles the actual rate of entropy decrease over time when the noise profile changes.
Lev: If we get that full understanding, we can move away from worst-case scenarios based on uniform assumptions and start designing protection schemes that are tailored to specific noise distributions.
Kai: They also suggest a complete classification of equality cases, which means figuring out exactly what kind of state makes the lower bound actually hit its minimum.
Mira: That would help us understand the precise conditions under which we can achieve that ideal one-channel structure they proved exists, without any extra room for error in our theory.
Lev: Understanding those equality cases is crucial because it tells us exactly when a simplified model like the single active two-level block really applies to the physical system.
Conclusion: Kai: So, to wrap up, we're talking about the paper "Sharp Entropy Cost of Coherent Leakage Across a Block Decomposition." The main point is that they've established a sharp entropy cost theorem and this logarithmic boundary law for removing cross-block coherence.
Mira: Basically, they show that this loss doesn't just vanish as you get better; it grows logarithmically with the leakage mass if you keep the floor high enough.
Lev: For error correction, that means we can set realistic limits on how much noise we can tolerate before the cost skyrockets.
Kai: And they’ve given us this one-channel principle, which says that for any given constraints on your system, the minimum loss always comes from just concentrating all the coherence into one single two-level block.
Mira: That idea is pretty powerful because it simplifies things; it suggests we don't need to optimize a huge number of blocks at once when minimizing this specific type of cost.
Lev: If that’s true, then for building real hardware, we can stop worrying about distributing coherence across many channels unnecessarily and just focus on optimizing that one effective qubit structure.
Kai: They also extended the core mechanism to systems with graded automorphisms and corner algebras, proving the principle is more robust than they first showed.
Mira: That extension shows this one-channel idea isn't just for simple models; it holds up in more mathematically complex settings, which is pretty significant for theory.
Lev: And their simulation routine for the uniform block dephasing orbit gives us a concrete way to check that the entropy loss rate actually follows the bound they derived.
Kai: So, this paper gives us a precise tool—a sharp entropy cost theorem—to measure and predict exactly how much information we lose when dephasing across blocks in quantum systems.
Mira: It's less about finding a perfect solution and more about understanding the fundamental limits of coherence protection under these block-dephasing maps.
Lev: It changes how we approach noise modeling for error correction; it gives us a concrete cost to calculate instead of just a vague estimate.
Kai: We’ll be looking at how this sharp cost relates to other transport studies next week, where we look at quantum transport along bosonic tight-binding chains connected to Markovian reservoirs.
Hassan Nasreddine
FortiSec Systems Inc.
quant-ph, math-ph, math.MP
Submitted: 2026-05-10
Updated: 2026-11-18
Comments: 27 pages, no figures. Accepted for publication in Journal of Mathematical Physics (AIP Publishing). Revised following peer review; principal mathematical results unchanged
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
Importance score: 92/100
The gist: Sharp Entropy Cost of Coherent Leakage Across a Block Decomposition is an investigation into quantifying the information-theoretic cost associated with removing cross-block coherence in quantum…
Key concepts
- Block Decomposition
- Splitting the density matrix H into P and Q parts (H = PH ⊕ QH) based on a superselection rule or measurement partition. This separates the system into protected blocks (P) and leakage blocks (Q), allowing researchers to track coherence specifically within these partitions.
- Block-Coherence Entropy
- The relative entropy D(ρ∥Πρ), which measures the informational distance between the original state ρ and its dephased shadow Πρ. It quantifies how much information is lost when cross-block coherence is removed, serving as the primary measure of the cost.
- Midpoint Bogoliubov–Kubo–Mori Coercivity
- A mathematical estimate proving a lower bound on the block-coherence entropy based on operator norms. It establishes that for certain conditions, this lower bound is strictly positive, providing a fundamental constraint on how small the coherence cost can be.
- Coherence Compression Principle
- A principle stating that any state with multiple coherent blocks can be simplified into one active block via a compression map without increasing entropy. This explains why the high-dimensional problem collapses to a simpler two-level structure at the minimum cost.
Terminology
Summary
Sharp Entropy Cost of Coherent Leakage Across a Block Decomposition is an investigation into quantifying the information-theoretic cost associated with removing cross-block coherence in quantum systems using block dephasing. p. 0 In quantum information, Π is the ideal block-dephasing
channel associated with this decomposition. The relative entropy
D(ρ∥Πρ) := Tr[ρ(logρ −logΠρ)]
is the block-coherence entropy: it measures the entropy cost of removing cross-block coherence
from ρ, or equivalently the informational distance from ρ to its dephased shadow. In the present two-block setting, they are derived from the midpoint
BKM operator-kernel bound and combined with the exact constrained one-channel variational
theorem. The main contribution of the present paper is the combination, in a single two-block
framework, of two independent ingredients: (i) a sign-symmetric midpoint BKM coercivity theorem,
which exploits the block-sign involution U = P−Q and the scalar resolvent inequality to establish
HM+tY (Y,Y) ≥ HM(Y,Y) for all t ∈ [0,1); and (ii) an exact constrained entropy minimization
theorem identifying a one-channel extremizer—a single active two-level block—that achieves the
minimum block-coherence entropy at prescribed support floor, leakage mass, and Hilbert–Schmidt
off-block coherence.Where the paper says it
Where the paper says it
Block Structure and Core Concepts
The paper introduces the block decomposition of a density matrix as H = PH ⊕ QH arising from a superselection rule or measurement partition. The map Π is defined as the block-dephasing map Πρ:= PρP ⊕ QρQ, which removes all cross-block coherence while preserving each block’s diagonal data. p. 0 In all these cases, the block-diagonal state space {σ ≥ 0 : Trσ = 1, σ =
PσP⊕QσQ} represents states with no quantum coherence across the boundary. Block dephasing is the map Πρ := PρP ⊕ QρQ that removes all cross-block coherence while
preserving each block’s diagonal data. In quantum information, Π is the ideal block-dephasing
channel associated with this decomposition. The relative entropy
D(ρ∥Πρ) := Tr[ρ(logρ −logΠρ)]
is the block-coherence entropy: it measures the entropy cost of removing cross-block coherence
from ρ, or equivalently the informational distance from ρ to its dephased shadow. Coherent leakage refers to the off-block component
B = PρQ, whose Frobenius norm squared c = ∥B∥
2
F we call the quadratic coherence. The leakage
mass is εQ = Tr(QρQ).Where the paper says it
Where the paper says it
Where the paper says it
Midpoint Coercivity and Extremal States
The authors prove a midpoint Bogoliubov–Kubo–Mori coercivity estimate, showing that the operator-kernel lower bound D(ρ∥Πρ) ≥ Tr[B∗omega−1A,C(B)] for strictly positive diagonal blocks. This is established using a block-sign involution U = P−Q to obtain a sign-symmetric midpoint lower bound for the BKM Hessian. p. 0 First, we prove
a midpoint Bogoliubov–Kubo–Mori coercivity estimate, giving the operator-kernel lower
bound D(ρ∥Πρ) ≥ Tr[B
∗Ω
−1
A,C
(B)] for strictly positive diagonal blocks. The proof uses
the block-sign involution U = P−Q to obtain a sign-symmetric midpoint lower bound for
the BKM Hessian. Second, we determine the exact minimum in the associated constrained
entropy-minimization problem: at fixed protected-sector floor, leakage mass, and quadratic
coherence, the minimum can be achieved by a single active two-level block, independently
of the ambient dimension. Third, under a uniform floor A ≥ a0P and 0 < εQ ≤ a0/2, we obtain the logarithmic boundary law D(ρ∥Πρ) ≥ ∥B∥
2
F
log(a0/εQ), showing that the lowerbound coefficient multiplying the quadratic coherence diverges logarithmically as leakage
mass tends to zero. The minimum can be attained by a single active two-level block: a state of the form
a∗
√
c √
c εQ
on one pair of vectors (|p1⟩,|q1⟩), with incoherent spectators at the floor value
a0 in the remaining P-directions. The one-channel principle states that the constrained minimum can always be realized by a twolevel configuration, regardless of the ambient dimension.Where the paper says it
Where the paper says it
Where the paper says it
Logarithmic Boundary Law and Spectral Decomposition
Under uniform floor assumptions A ≥ a0P and 0 < εQ ≤ a0/2, the paper derives the logarithmic boundary law D(ρ∥Πρ) ≥ ∥B∥2F log(a0/εQ). This shows that the lower-bound coefficient multiplying the quadratic coherence diverges logarithmically as leakage mass tends to zero. p. 0 Second, we determine the exact minimum in the associated constrained
entropy-minimization problem: at fixed protected-sector floor, leakage mass, and quadratic
coherence, the minimum can be achieved by a single active two-level block, independently
of the ambient dimension. Third, under a uniform floor A ≥ a0P and 0 < εQ ≤ a0/2, we obtain the logarithmic boundary law D(ρ∥Πρ) ≥ ∥B∥
2
F
log(a0/εQ), showing that the lowerbound coefficient multiplying the quadratic coherence diverges logarithmically as leakage
mass tends to zero. The
logarithmic bound is asymptotically sharp on the two-level family (Proposition III. Spectral channel sensitivity For B ̸= 0, the BKM bound decomposes as
Tr[B
∗Ω
−1
A,C
(B)] = ∥B∥
2
F ∑
α,β
wαβ L(aα,cβ
),
where wαβ = |⟨eα,B fβ
⟩|2/∥B∥
2
F
and L(a,c) = log(a/c)/(a−c) is the BKM logarithmic mean kernel (Proposition III. The full operator bound detects which (A,C)-eigenbasis channels are active;
coarse scalar data (a0, εQ,c) do not determine the BKM cost. The operator–scalar separation theorem (Proposition III.Where the paper says it
Where the paper says it
Where the paper says it
Variational Problem and Compression Principle
The paper solves the constrained extremal problem by proving that the infimum D(ρ∥Πρ) = Φ(a∗, εQ,c), where a∗ = 1−εQ −(dP −1)a0. The Coherence Compression Principle (Theorem IV. p. 1 Thus the high-dimensional constrained
problem collapses to a scalar 2×2 problem. The Coherence Compression Principle (Theorem IV. The technical mechanism is the Coherence
Compression Principle (Theorem IV. 5): any state with multiple coherent blocks can be mapped by
a CPTP compression to one with a single active block, without increasing entropyWhere the paper says it
Where the paper says it
Extensions and Physical Interpretation
The midpoint BKM mechanism extends to finite-dimensional trace-preserving Z2-graded automorphisms (Theorem VII.1) and under bounded strict-positivity assumptions, to finite-trace corners of semifinite von Neumann algebras (Theorem VII. p. 2 The midpoint BKM coercivity mechanism extends to general finite-dimensional
Z2-graded automorphisms (Theorem VII. 1) and, under bounded strict-positivity assumptions, to
finite-trace corners of semifinite von Neumann algebras (Theorem VII. The minimum can be attained by a single active two-level block: a state of the form
a∗
√
c √
c εQ
on one pair of vectors (|p1⟩,|q1⟩), with incoherent spectators at the floor value
a0 in the remaining P-directions. The one-channel principle states that the constrained minimum can always be realized by a twolevel configuration, regardless of the ambient dimension. The BKM estimate also yields a dissipation lower bound
along an idealized uniform block-dephasing orbit. Finally, the midpoint mechanism extends to finite-dimensional trace-preserving Z2-graded automorphisms and bounded finitetrace semifinite corners.Where the paper says it
Where the paper says it
Where the paper says it
Conclusion and Limitations
The main contributions are the sharp entropy-cost theorem, the BKM midpoint coercivity and logarithmic boundary law, and the auxiliary extension of the midpoint mechanism to graded settings. The paper notes limitations, such as not treating non-uniform dephasing rates or full modified logarithmic Sobolev inequalities. p. 2 1) and, under bounded strict-positivity assumptions, to
finite-trace corners of semifinite von Neumann algebras (Theorem VII. We do not prove a full
modified logarithmic Sobolev inequality, do not treat non-uniform dephasing rates, do not establish a theory for unbounded densities or type III von Neumann algebras (these require additional
modular-theoretic arguments), and do not give a complete equality classification for the variational
minimizers. 1) and, under bounded strict-positivity assumptions, to
finite-trace corners of semifinite von Neumann algebras (Theorem VII. We do not prove a full
modified logarithmic Sobolev inequality, do not treat non-uniform dephasing rates, do not establish a theory for unbounded densities or type III von Neumann algebras (these require additional
modular-theoretic arguments), and do not give a complete equality classification for the variational
minimizers.Where the paper says it
Where the paper says it
Appendix Details
The appendix provides the explicit SVD pinching CPTP map E(X) and the exact spectral formula for the BKM kernel Tr[B∗omega−1A,C(B)] = ∑α,β ⟨eα,B fβ⟩2L(aα,cβ). The proof of the Polygon Modulus Lemma is also included in Appendix B. p. 4 Section VIII concludes. Appendices contain the SVD pinching construction, the polygon
modulus lemma, boundary cases for the Coherence Compression Principle, and the Petz-metric
extension. (Kubo–Ando concavity of operator means is used separately in Appendix D for the Petz-metric analogue; it is not part of the BKM argument. ) Integrating this midpoint lower bound along the affine
path ρ(t) = M +tY from t = 0 to t = 1 yields the BKM operator-kernel lower bound (Theorem III. Together:
the BKM operator controls the exact leading asymptotic term; the logarithmic corollary therefore
captures the same leading order. Spectral channel formula and operator–scalar separation
Proposition III.Where the paper says it
Where the paper says it
Where the paper says it
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
-
Implement an entropy certificate for coherent leakage from a protected subspace by measuring PrhoQ, which distinguishes it from block-diagonal states with the same leakage mass epsilon Q but zero PrhoQ.
-
Develop a method to evaluate the BKM bound Tr[B∗Omega−1A,C(B)] using spectral channel representations, specifically decomposing it into Tr[B∗Omega−1A,C(B)] = sum alpha,beta walpha beta L(aα,cβ), where L is the BKM logarithmic mean kernel.
-
Create a diagnostic tool that uses the scalar logarithmic law D(ρ∥Πρ) ≥ B2F log a0/epsilon Q to quantify how the entropy cost per unit Hilbert-Schmidt off-block coherence grows logarithmically as leakage mass tends to zero under a uniform floor A ≥ a0P.
-
Design an optimizer for constrained entropy minimization that utilizes the Coherence Compression Principle, showing that the minimum block-coherence entropy D(ρ∥Πρ) equals Φ(a∗, epsilon Q, c) and is attained by a single active two-level block regardless of ambient dimension.
-
Build a simulation routine to test the one-channel principle by verifying that for any state with fixed floor A >= a0P, leakage mass epsilon Q, and quadratic coherence c = PrhoQ2F, the entropy cost is minimized when coherence is concentrated into one effective two-level channel.
-
Extend the midpoint BKM mechanism to systems described by finite-dimensional trace preserving Z2 graded automorphisms and to finite trace corners of semifinite von Neumann algebras under bounded strict positivity assumptions.
-
Create a computational framework for simulating an idealized uniform block dephasing orbit, demonstrating that the dissipation rate satisfies-d/dt D(ρt∥Πρ) ≥ 2G e(-2Gt) Tr[B∗Omega−1A,C(B)] when A >= a0P and epsilon Q ≤ a0/2.
Abstract
Let H=P H plus Q H and let Πρ=PρP plus QρQ be the two-block pinching of a density matrix. We study the block-coherence entropy D(ρ|Πρ), the relative-entropy cost of removing cross-block coherence. First, we prove a midpoint Bogoliubov--Kubo--Mori (BKM) coercivity estimate, giving the operator-kernel lower bound D(ρ|Πρ) at least Tr[B*Ω A,C-1(B)] for strictly positive diagonal blocks. The proof uses the block-sign involution U=P-Q to obtain a sign-symmetric midpoint lower bound for the BKM Hessian. Second, we solve the associated constrained entropy-minimization problem: at fixed protected-sector floor, leakage mass, and quadratic coherence, the minimum can be achieved by a single active two-level block, independently of the ambient dimension. Third, under a uniform floor A at least a 0P and 0< epsilon Q at most a 0/2, we recover the logarithmic boundary law D(ρ|Πρ) at least|B| F 2 (a 0/epsilon Q), showing a divergent entropy cost per unit quadratic coherence as leakage mass tends to zero. As a finite-dimensional application, the estimates give an entropy certificate for coherent leakage from a protected subspace: PρQ measures protected--leakage coherence, and the constrained entropy minimum is attained by a state supported on a single effective two-level transition. The BKM estimate also yields a dissipation lower bound along an idealized uniform block-dephasing orbit. Finally, the midpoint mechanism extends to finite-dimensional trace-preserving Z 2-graded automorphisms and finite von Neumann algebras under bounded uniform positivity assumptions.
Sources
- Quantifying Superposition
- Entropy Moduli and Support-Sensitive BKM Coercivity for Rank-Deficient Non-Commutative Markov Semigroups
Related papers
- Computational regimes in matrix-product-state-based quantum trajectory simulations
- Toward chemically accurate quantum simulations on quantum hardware: A quantum-HPC hybrid approach
- Transversal surface-code game with reconfigurable qubits
- Fulqrum: a generalized framework for quantum subspace diagonalization
- Flux-Activated Resonant Control of a Bosonic Quantum Memory
- Machine learning the effects of many quantum measurements