pith. sign in

arxiv: 2303.18224 · v2 · pith:GZHBR4QFnew · submitted 2023-03-31 · 🪐 quant-ph · math-ph· math.FA· math.MP

Quantum Thermal State Preparation

Chi-Fang Chen , Michael J. Kastoryano , Fernando G.S.L. Brand\~ao , Andr\'as Gily\'en This is my paper

Pith reviewed 2026-05-19 13:07 UTC · model grok-4.3

classification 🪐 quant-ph math-phmath.FAmath.MP
keywords quantum Gibbs samplingthermal state preparationLindbladianquantum simulationthermal field doubleMarkov chain Monte Carloquantum master equation
0
0 comments X p. Extension
pith:GZHBR4QF Add to your LaTeX paper What is a Pith Number? 
\usepackage{pith}
\pithnumber{GZHBR4QF}

Prints a linked pith:GZHBR4QF badge after your title and writes the identifier into PDF metadata. Compiles on arXiv with no extra files. Learn more

The pith

Continuous-time quantum algorithms prepare thermal states by simulating natural Lindblad master equations.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that earlier Monte Carlo approaches to preparing quantum thermal states ran into fundamental obstacles from energy-time uncertainty, but simple continuous-time samplers can bypass them. These samplers work by directly simulating quantum master equations (Lindbladians) that mimic how real physical systems reach equilibrium. The authors also give the first provably efficient method for preparing purified thermal states, called thermal field double states, for systems whose Lindbladians have a large spectral gap, and they obtain an extra quantum-walk speedup. Their cost bounds depend explicitly on temperature, desired accuracy, and the Lindbladian mixing time. They supply the first rigorous non-asymptotic proof that physically motivated Lindbladians thermalize in finite time under approximate detailed balance.

Core claim

We introduce simple continuous-time quantum Gibbs samplers that overcome these obstacles by efficiently simulating Nature-inspired quantum master equations (Lindbladians). In addition, we construct the first provably accurate and efficient algorithm for preparing certain purified Gibbs states (called thermal field double states in high-energy physics) of rapidly thermalizing systems; this algorithm also benefits from a quantum walk speedup. Our algorithms' costs have a provable dependence on temperature, accuracy, and the mixing time (or spectral gap) of the relevant Lindbladian. We complete the first rigorous proof of finite-time thermalization for physically derived Lindbladians bydevelops

What carries the argument

Continuous-time quantum Gibbs samplers that simulate Lindbladian master equations with approximate detailed balance and a large spectral gap.

If this is right

  • Thermal-state preparation cost becomes polynomial in system size, inverse temperature, and mixing time for qualifying systems.
  • Purified thermal field double states become preparable with an extra quantum-walk speedup.
  • Finite-time thermalization is now rigorously established for a broad class of physically derived Lindbladians.
  • Quantum simulation workflows can now import the practical success of classical Markov-chain Monte Carlo methods.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same Lindbladian framework may extend to preparing other equilibrium states beyond thermal ones if similar spectral-gap conditions hold.
  • High-energy physics simulations that rely on thermal field doubles could become feasible on near-term quantum hardware for rapidly thermalizing models.
  • Classical MCMC intuition about mixing times now has a direct quantum counterpart that can be benchmarked on small systems.

Load-bearing premise

The systems must be rapidly thermalizing so that the relevant Lindbladian has a spectral gap large enough for polynomial preparation cost, and the Lindbladians must admit an efficient implementation together with approximate detailed balance.

What would settle it

A concrete physical system whose Lindbladian mixing time grows exponentially with system size while still satisfying the rapid-thermalization assumption, or an explicit counterexample where the nonasymptotic secular approximation fails to preserve detailed balance.

read the original abstract

Preparing ground states and thermal states is essential for simulating quantum systems on quantum computers. Despite the hope for practical quantum advantage in quantum simulation, popular state preparation approaches have been challenged. Monte Carlo-style quantum Gibbs samplers have emerged as an alternative, but prior proposals have been unsatisfactory due to technical obstacles rooted in energy-time uncertainty. We introduce simple continuous-time quantum Gibbs samplers that overcome these obstacles by efficiently simulating Nature-inspired quantum master equations (Lindbladians). In addition, we construct the first provably accurate and efficient algorithm for preparing certain purified Gibbs states (called thermal field double states in high-energy physics) of rapidly thermalizing systems; this algorithm also benefits from a quantum walk speedup. Our algorithms' costs have a provable dependence on temperature, accuracy, and the mixing time (or spectral gap) of the relevant Lindbladian. We complete the first rigorous proof of finite-time thermalization for physically derived Lindbladians by developing a general analytic framework for nonasymptotic secular approximation and approximate detailed balance. Given the success of classical Markov chain Monte Carlo (MCMC) algorithms and the ubiquity of thermodynamics, we anticipate that quantum Gibbs sampling will become indispensable in quantum computing.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript introduces continuous-time quantum Gibbs samplers that simulate Lindbladian master equations to prepare thermal states on quantum computers, overcoming energy-time uncertainty issues in prior Monte Carlo approaches. It also presents the first provably accurate and efficient algorithm for preparing thermal field double (purified Gibbs) states of rapidly thermalizing systems, benefiting from a quantum walk speedup. A key contribution is a general analytic framework for nonasymptotic secular approximation and approximate detailed balance, enabling the first rigorous proof of finite-time thermalization for physically derived Lindbladians. Algorithmic costs are proven to depend on temperature, accuracy, and the mixing time or spectral gap.

Significance. If the results hold, this work significantly advances quantum algorithms for state preparation in quantum simulation by providing efficient, provably correct methods with explicit parameter dependence, analogous to the role of MCMC in classical computing. The rigorous analytic framework for nonasymptotic secular approximation, the explicit dependence of costs on temperature, accuracy, and spectral gap, and the quantum walk speedup for TFD states are notable strengths that should be credited.

major comments (2)
  1. [Analytic framework for nonasymptotic secular approximation and approximate detailed balance] The section developing the general analytic framework for nonasymptotic secular approximation: the error analysis does not supply explicit non-asymptotic bounds on the integral of the residual bath correlation functions that scale independently of the number of modes or the ultraviolet cutoff. This is load-bearing for the central claim of a general rigorous proof of finite-time thermalization for physically derived Lindbladians, since without such control the secular approximation error may fail to remain O(1/poly(n,1/ε,1/gap)) for broad classes of microscopic system-bath models.
  2. [Algorithm for purified Gibbs states (thermal field double states)] The construction and analysis of the algorithm for thermal field double states: the claimed polynomial cost and quantum walk speedup rest on the assumption that the relevant Lindbladian in the purified space inherits a sufficiently large spectral gap from the original system; the manuscript should explicitly verify that the gap lower bound carries over without additional factors that could degrade the overall scaling.
minor comments (2)
  1. [Abstract] The abstract could more explicitly quantify the polynomial dependence on the spectral gap to highlight the improvement over prior work.
  2. [Preliminaries and notation] Notation for the Lindbladian generators and the interaction picture transformation would benefit from a brief concrete example of a system-bath Hamiltonian to aid readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their careful reading of the manuscript and for the constructive comments. We are glad that the referee recognizes the potential significance of the continuous-time quantum Gibbs samplers and the rigorous framework for finite-time thermalization. We address each major comment below and will revise the manuscript to incorporate additional clarifications and explicit statements as needed.

read point-by-point responses

  1. Referee: [Analytic framework for nonasymptotic secular approximation and approximate detailed balance] The section developing the general analytic framework for nonasymptotic secular approximation: the error analysis does not supply explicit non-asymptotic bounds on the integral of the residual bath correlation functions that scale independently of the number of modes or the ultraviolet cutoff. This is load-bearing for the central claim of a general rigorous proof of finite-time thermalization for physically derived Lindbladians, since without such control the secular approximation error may fail to remain O(1/poly(n,1/ε,1/gap)) for broad classes of microscopic system-bath models.

    Authors: We thank the referee for emphasizing the importance of explicit control over the secular approximation error. The manuscript develops the nonasymptotic framework in the relevant technical section by expressing the error in terms of the time-integrated residual bath correlation functions and deriving bounds under the assumption of standard physical decay properties (e.g., exponential or faster decay for finite-bandwidth baths). For broad classes of microscopic models, these integrals remain bounded by constants independent of the number of modes and ultraviolet cutoff. To strengthen the presentation and directly address the concern, we will add an explicit lemma in the revised version that states the non-asymptotic bound on the integrated residual correlations and confirms that the overall secular error scales as O(1/poly(n,1/ε,1/gap)) for the relevant models. This addition will make the central claim fully rigorous without changing the main results. revision: partial

  2. Referee: [Algorithm for purified Gibbs states (thermal field double states)] The construction and analysis of the algorithm for thermal field double states: the claimed polynomial cost and quantum walk speedup rest on the assumption that the relevant Lindbladian in the purified space inherits a sufficiently large spectral gap from the original system; the manuscript should explicitly verify that the gap lower bound carries over without additional factors that could degrade the overall scaling.

    Authors: We appreciate the referee's request for explicit verification of the spectral gap in the purified space. In the construction, the purified Lindbladian is defined symmetrically on the system-ancilla space such that its generator preserves the essential mixing properties of the original system Lindbladian. The spectral gap of the purified generator is at least a constant factor (specifically 1/2) times the original gap, arising from the tensor-product-like structure with the identity component on the ancilla; this constant factor does not affect the polynomial scaling in 1/gap. We will revise the manuscript to include an explicit remark or short lemma verifying this inheritance and the constant factor, thereby confirming that the polynomial cost and quantum walk speedup remain intact. revision: yes

Circularity Check

0 steps flagged

Derivation chain self-contained; no reduction to fitted inputs or self-citation

full rationale

The paper's central results are explicit algorithms whose runtime is bounded by the mixing time/gap of the Lindbladian plus temperature and accuracy parameters, together with a new analytic framework for nonasymptotic secular approximation. These bounds are stated directly in terms of the assumed spectral gap and do not reduce, via any equation in the manuscript, to a quantity that is itself computed or fitted inside the paper. The secular-approximation framework is developed from first principles within the present work rather than imported via self-citation as a load-bearing uniqueness theorem. No step equates a derived prediction to its own input by construction, and the efficiency claims remain parametrically dependent on externally verifiable quantities (gap, bath correlation decay) that are not redefined inside the derivation.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claims rest on standard domain assumptions from open quantum systems theory rather than new free parameters or invented entities.

axioms (1)
  • domain assumption The relevant Lindbladians satisfy approximate detailed balance and admit efficient implementation on a quantum computer.
    Invoked to guarantee that the continuous-time simulation yields the desired thermal state with cost polynomial in the inverse gap and temperature.

pith-pipeline@v0.9.0 · 5754 in / 1347 out tokens · 58725 ms · 2026-05-19T13:07:44.377772+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 20 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Accelerating quantum Gibbs sampling without quantum walks

    quant-ph 2026-04 unverdicted novelty 8.0

    A factorization of the parent Hamiltonian into noncommutative first-order operators enables a walk-free QSVT algorithm with quadratic gap improvement for preparing purified Gibbs states under exact KMS detailed balance.

  2. Computing the free energy of quantum Coulomb gases and molecules via quantum Gibbs sampling

    quant-ph 2026-04 unverdicted novelty 8.0

    Quantum algorithm with rigorous truncation error bounds and spectral gap guarantees for free energy estimation in finite-temperature Coulomb quantum systems via Markovian Gibbs sampling.

  3. Simulating Thermal Properties of Bose-Hubbard Models on a Quantum Computer

    quant-ph 2026-04 unverdicted novelty 8.0

    A new rigorous Gibbs sampling method is given for bosonic models by proving that their dissipative generators have positive spectral gaps, enabling efficient quantum preparation of thermal states for Bose-Hubbard Hami...

  4. Typical Mixing and Rare-State Bottlenecks in Open Quantum Systems

    quant-ph 2026-05 unverdicted novelty 7.0

    Typical trace-distance relaxation concentrates around a mean in open quantum systems, producing typical mixing times separated from worst-case by rare-state bottlenecks that scale logarithmically, linearly, or exponen...

  5. Overcoming the Lamb Shift in System-Bath Interaction Models via KMS Detailed Balance: High-Accuracy Thermalization with Time-Bounded Interactions

    quant-ph 2026-04 unverdicted novelty 7.0

    KMS detailed balance on the jump rates of an approximate Lindbladian overcomes arbitrary Lamb shifts to produce high-accuracy Gibbs states with O(ε^{-1}) mixing time.

  6. Preparing High-Fidelity Thermofield Double States

    quant-ph 2026-05 unverdicted novelty 6.0

    A gapped parent Hamiltonian built from two copies of a target Hamiltonian plus ultra-local inter-copy couplings allows adiabatic preparation of high-fidelity thermofield double states for ETH-obeying systems.

  7. Note on Strong Quantum Markov Properties

    quant-ph 2026-05 unverdicted novelty 6.0

    The strong Markov property holds if and only if the state has correlation decay for suitable observables, enabling single-copy multi-observable estimation and forcing local marginals of such states to be close or well...

  8. Overcoming the Lamb Shift in System-Bath Interaction Models via KMS Detailed Balance: High-Accuracy Thermalization with Time-Bounded Interactions

    quant-ph 2026-04 unverdicted novelty 6.0

    Proves KMS detailed balance on the transition part of an approximate Lindbladian suffices for the fixed point to approach the Gibbs state arbitrarily closely regardless of Lamb shift structure, giving O(ε^{-1}) therma...

  9. Rapid mixing for high-temperature Gibbs states with arbitrary external fields

    quant-ph 2026-04 unverdicted novelty 6.0

    High-temperature Gibbs states with arbitrary external fields admit O(log n) quantum mixing via a detailed-balance Lindbladian and exhibit classical sampling hardness for β < 1.

  10. Quantum Gibbs sampling through the detectability lemma

    quant-ph 2026-04 conditional novelty 6.0

    Detectability lemma enables Gibbs sampling without Lindbladian simulation, yielding O(M) cost reduction for M-term local Lindbladians and quadratic speedup in spectral gap for frustration-free and commuting cases.

  11. Thermalization from quenching in coupled oscillators

    quant-ph 2025-12 conditional novelty 6.0

    Sudden quenches in a pair of coupled oscillators produce exact or approximate thermalization of a quantum harmonic oscillator to arbitrary temperatures via solvable equations on the Gaussian covariance matrix.

  12. Variational Thermal State Preparation on Digital Quantum Processors Assisted by Matrix Product States

    quant-ph 2025-10 unverdicted novelty 6.0

    A variational framework assisted by matrix product states prepares approximate thermal Gibbs states for 1D lattices up to 30 sites and 2D lattices up to 6x6 using up to 44 qubits, with a demonstration on IBM Heron hardware.

  13. Quantum algorithms based on quantum trajectories

    quant-ph 2025-09 unverdicted novelty 6.0

    Quantum trajectory algorithm achieves additive O(T + log(1/ε)) query complexity for simulating dissipative Lindbladians.

  14. Adiabatic preparation of thermal states and entropy-noise relation on noisy quantum computers

    quant-ph 2025-09 conditional novelty 6.0

    Adiabatic evolution prepares local thermal states from initial Gibbs states while conserving entropy density in the thermodynamic limit, with mirror-circuit benchmarking of hardware noise entropy demonstrated experime...

  15. Constrained free energy minimization for the design of thermal states and stabilizer thermodynamic systems

    quant-ph 2025-08 unverdicted novelty 6.0

    Benchmarks gradient-ascent algorithms for constrained free energy minimization on quantum Heisenberg models and stabilizer codes, with applications to thermal state design and fixed-temperature quantum encoding.

  16. Planckian bound on quantum dynamical entropy

    quant-ph 2025-07 unverdicted novelty 6.0

    A simplified version of quantum dynamical entropy is introduced, its growth rate is computed from correlation functions in the thermodynamic limit, and a Planckian bound on the rate is conjectured.

  17. Efficient Quantum Gibbs Sampling with Local Circuits

    quant-ph 2025-06 unverdicted novelty 6.0

    Introduces local-circuit approximations to quasilocal dissipative processes for efficient, provably convergent quantum Gibbs sampling at high temperatures.

  18. Quantum Markov chain Monte Carlo method with programmable quantum simulators

    quant-ph 2025-05 unverdicted novelty 6.0

    A quantum MCMC algorithm leveraging the MBL phase and its thermal-to-localized transition to tune acceptance rates and sample thermal distributions on programmable quantum simulators for combinatorial optimization.

  19. How to Build a Quantum Supercomputer: Scaling from Hundreds to Millions of Qubits

    quant-ph 2024-11 accept novelty 4.0

    A comprehensive review of scaling paths for superconducting quantum computers, with resource and sensitivity analyses for utility-scale applications under realistic error distributions.

  20. Quantum Computing Beyond Ground State Electronic Structure: A Review of Progress Toward Quantum Chemistry Out of the Ground State

    physics.chem-ph 2025-09 unverdicted novelty 2.0

    Review of quantum computing methods and potential for non-ground-state quantum chemistry including reaction dynamics, mechanisms, and finite temperatures.

Reference graph

Works this paper leans on

12 extracted references · 12 canonical work pages · cited by 19 Pith papers

  1. [1]

    polarization

    Bootstrapping the secular approximation Lemma B.4(Bootstrapping the secular approximation). Consider the decompositionf = ∑ j∈Jfj of the weight function, wherefj : St0→C, and letµj := min { µ≥0:  ˆfj(ω)·1 (|ω|>µ)  = 0 } . If the Hamiltonian ¯H has discretized spectrum so thatB ⊂ω0Z, βµj ≤1, max¯ω∈B,j∈J ¯ω+µj =: ν∈Sω0, γ: Sω0 →R+ is such that γ(¯ω)/γ...

    work page

  2. [2]

    discretized

    Fourier Transform with uniform weights For simpler implementation, we can also work with the Fourier Transform with uniform weight (which is not smooth), leading to slightly worse bounds than the Gaussian damped case of Theorem I.3. Theorem B.1(Uniform weight for Fourier Transform). Consider the discriminant proxyDβ(3.6) with the plain Fourier Transform ˆ...

    work page

  3. [3]

    finitie-time

    (C3) Also observe that due to (C2) we have for allω∈R that ∑ ¯t∈S⌈N⌋ t0 t0 fT (t)e−iωt−fT (¯t)e−iω¯t  [¯t,¯t+t0) ≤ ∑ ¯t∈S⌈N⌋ t0 t0 fT (t)e−iωt−fT (¯t)e−iωt [¯t,¯t+t0) +t0 fT (¯t)e−iωt−fT (¯t)e−iω¯t  [¯t,¯t+t0) ≤ ∑ ¯t∈S⌈N⌋ t0 t0 fT (t)−fT (¯t)  [¯t,¯t+t0) +t0|fT (¯t)| e−iωt−e−iω¯t  [¯t,¯t+t0) ≤ ∑ ¯t∈S⌈N⌋ t0 t0 fT (t)−fT (¯t)  ...

    work page

  4. [4]

    Proposition D.1 (Approximate detailed balance)

    Bounds for approximate detailed balance Here, we show that approximate detailed balance for the Lindbladian of interestLsec. Proposition D.1 (Approximate detailed balance). Suppose the secular approximation forL(CGME ) (D4) is truncated at energyµ. Then, 1 2 D(ρ,Lsec)−D(ρ,Lsec)† 2−2≤O ( β √µ T ) . Proof. We simply telescope by inserting the algorithmi...

    work page

  5. [5]

    The resulting bounds now depend on two mixing times, and we do not have a desirable conversion between the two mixing times

    Effects of the Lamb-shift term In this section, we include the unitary part of the CGME generator. The resulting bounds now depend on two mixing times, and we do not have a desirable conversion between the two mixing times. Still, one can upper bound both via the spectral gap of the Hermitian part of the dissipative partHdiss. Theorem D.2 (Fixed point of ...

    work page

  6. [6]

    Altogether: Proof of fixed point correctness (Theorem D.2) We now put together the error bounds for the full CGME Lindbladian. 63 Proof of Theorem D.2.Recall the bound on the fixed point error ∥ρfix(L)−ρ∥1 =∥ρfix(L)−ρfix(Lsec)∥1 +∥ρfix(Lsec)−ρfix(L′)∥1 ≤O ( ∥L−Lsec∥1−1tmix(L) +∥D(ρ,Lsec)−D(ρ,L′)∥2−2 ς−2(D(ρ,Lsec)) ) ≤O ( ∥L−Lsec∥1−1tmix(L) +∥Ldiss−Ldiss,s...

    work page

  7. [7]

    Unfortunately, the sharp truncation from Section A1 does not seem to work here because the truncation error is related to the 1-norm∥f∥1 (instead of 2-norm∥f∥2)

    Proof for secular approximation for time average (Lemma D.2) Intuitively, we want to truncate the Bohr frequency far from zero. Unfortunately, the sharp truncation from Section A1 does not seem to work here because the truncation error is related to the 1-norm∥f∥1 (instead of 2-norm∥f∥2). The 1-norm is more delicate to handle, forcing us tosmoothly trunca...

    work page

  8. [8]

    Proposition E.1(Bauer-Fike Theorem with multiplicity, cf

    Perturbation bounds for eigenvalues and eigenvectors In this section, we present some useful bounds for eigenvalue and eigenvector perturbation. Proposition E.1(Bauer-Fike Theorem with multiplicity, cf. [Bha97, Theorem VI.3.3 & Problem VI.8.6]). Perturb a normal matrixN by an arbitrary matrixA. Then, the spectrum ofN andN +A are∥A∥-close to each other: Sp...

    work page

  9. [9]

    Approximate detailed balance implies approximately correct fixed point When detailed balance holds approximately forρ, we still expect the fixed point to beapproximatelyρ; we provide a proof of this in this section, which relies on the matrix perturbation results (Section E1). Recall that in Section IIA, we defined the Hermitian and anti-Hermitian parts (...

    work page

  10. [10]

    sufficiently

    Perturbation bounds for Lindbladians regarding gaps and mixing times This section provides proof for scattered statements circling spectral gaps and mixing time. Most results are standard, except maybe the most technical result (Lemma E.2). Lemma II.1(Fixed point difference). For any two LindbladiansL1 andL2, the difference of their fixed points (in the S...

    work page

  11. [11]

    (G3) Proof

    (G2) Letδβ:=βj+1−βj, then the consecutive overlaps are large ⏐⏐⟨λ1(Dβj)|λ1(Dβj+1)⟩ ⏐⏐2 ≥7 10−O((δβ)2∥H2e−δβH∥). (G3) Proof. Let us evaluate the overlap between the ideal Gibbs states and rewrite using the Hilbert-Schmidt inner product ⏐⏐⟨√ρβj|√ρβj+1⟩ ⏐⏐2 = Tr [ e−βjH/2e−βjH/2e−δβH/2]2 Tr[e−βjH]Tr[e−βjHe−δβH] = ⟨e−δβH/2⟩2 βj ⟨I⟩βj⟨e−δβH⟩βj = 1−O((δβ)2⟨H2e−...

    work page

  12. [12]

    boosted shift-invariant in place phase estimation

    A simple lower bound onβdependence In this section, we prove a simple lower-bound for the temperature dependence in the sense of implementing a reflection about the purified Gibbs state. Proposition G.4(Lower-bound on simulation time). A circuit implementing the reflection operator Rβ,H :=I− ⏐⏐√ρβ,H ⟩ ⟨√ρβ,H ⏐⏐ using Hamiltonian simulation forH as a black...

    work page