{"id":"2f683318-46a0-42e4-b443-1ffd2cd7a7ac","arxiv_id":"2607.16857","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A quantum algorithm framework that converts any circuit-prepared state into an initial Hamiltonian for adiabatic state preparation, with numerical evidence that same-phase MPS warm starts improve adiabatic gaps.","lead":"The paper presents UPHAWS, a protocol that uses the Feynman–Kitaev clock Hamiltonian as a universal parent Hamiltonian to initialize adiabatic state preparation from any circuit-preparable state. It shows numerically that matrix-product-state warm starts can keep adiabatic gaps open in small model systems.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Momentum-truncation conjecture is unverified; H6 gap results, including the factor-of-two claim, may be truncation artifacts.","rationale":"The reader's CONDITIONAL verdict is appropriate. I re-read the protocol, the gap analysis, and the numerical sections. The clock-Hamiltonian gap scaling Θ(D_c^-2) is standard, and the warm-start construction is coherent. However, the H6 demonstration — the only example supporting the factor-of-two claim — rests entirely on the momentum-truncation conjecture stated in Sec. III D. The paper itself flags this assumption as a conjecture, and the benchmark in Fig. 4 is limited to n≤4 and T≤80, with no truncation-error estimate for H6. Thus the weakest assumption identified by the reader is also the one I would stress. I would not change the verdict: the conceptual construction stands, and a positive result on the convergence test would support acceptance, while a negative result would move the numerical claims closer to unverified.","tokens_in":18312,"tokens_out":30817,"duration_ms":300524,"concrete_test":"Recompute the H6 R=4.0 Å adiabatic path in the rotating frame with kcut fractions 10%, 20%, 30%, and 40% of T+1, and if feasible include a second-order perturbative estimate of the discarded-mode contribution using the resolvent of the retained block. Track the absolute minimum gap for the HF, MPS-D2, and MPS-D4 protocols. If the MPS-D4 minimum gap shifts by more than about 10% or the ordering relative to HF changes, the momentum-truncation conjecture is the load-bearing point of failure; if the values stabilize, the concern does not land.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The numerical support for the central H6 claim (Fig. 5: MPS-D4 warm start keeps min rescaled gap above 0.87, HF drops to 0.27) is computed in a truncated clock-momentum basis. The validity of this truncation is explicitly a conjecture in Sec. III D: off-diagonal couplings to discarded fast clock modes are assumed small relative to O(1) fast-mode energies. The stated prefactors are O(1/T) from the initialization penalty and O(D_c/T) from the rotated target operators. The second prefactor is not guaranteed small in the regime actually used: the H6 runs pad the clock only enough for 0.86 overlap with |+>_clock ⊗ |MPS> (Fig. 5 caption), and for a typical intermediate overlap not close to 1 this permits D_c/T ~ O(1). The truncation benchmark in Sec. IV B covers only GHZ systems with n≤4 and T≤80; no truncation-error estimate is reported for H6 (12 qubits, larger T and D_c). If the slow-fast couplings are not small, the truncated low-energy spectrum can differ from the true gap, so the advertised factor-of-two improvement could be an artifact of the truncation rather than of the warm start.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces UPHAWS, a protocol for warm-starting adiabatic state preparation (ASP) using the Feynman–Kitaev clock Hamiltonian as a universal parent Hamiltonian for any state with a known preparation circuit. The history state of the clock Hamiltonian is used as the initial state; by padding the clock, it approaches |+>_clock⊗|ψ_target>. The authors derive the initial spectral gap scaling Θ(D_c^{-2}) for deterministic circuits, extend the analysis to probabilistic circuits via amplitude amplification, and introduce a momentum-space truncation for classical simulation. They benchmark the method on a Z2-symmetric MPS family interpolating to a GHZ target and on the linear H6 chain under symmetric bond stretching, reporting that a bond-dimension-4 MPS warm start keeps the minimum rescaled gap above 0.87 while Hartree–Fock falls to 0.27.","tokens_in":18638,"tokens_out":13578,"duration_ms":126999,"significance":"If correct, UPHAWS provides a general and conceptually clean way to convert any circuit-preparable ansatz into an ASP initial state, with a concrete O(D_c^{-2}) gap cost. This would address limitations of local ansätze associated with the orthogonality catastrophe and offers a phase-aligned strategy for avoiding first-order transitions. The gap derivation is explicit and self-contained, and the numerical comparison is well motivated. However, the central numerical claims currently rest on an unverified truncation conjecture and on a rescaled-gap comparison that may not reflect the actual adiabatic runtime. These issues are fixable but require substantive additional work.","major_comments":[{"comment":"The reported 'factor of two' improvement is for the minimum gap rescaled by the gap at s=0. This rescaling hides the cost of the clock register. For the clock protocols Δ(0)=Θ(T^{-2}); with the 0.86-overlap padding this is orders of magnitude smaller than the HF initial gap. Thus the absolute minimum gap of the MPS-D4 protocol is likely much smaller than that of HF, not larger. The adiabatic runtime is governed by the absolute gap (and ||dH/ds||), so the rescaled quantity does not support the abstract's factor-of-two claim. Please report absolute gaps or explicit runtime estimates for all three protocols.","section":"Sec. IV C, Fig. 5, abstract"},{"comment":"The momentum-space truncation is load-bearing for the H6 results, but the smallness of the slow–fast off-diagonal couplings is explicitly left as a conjecture ('we conjuncture that both will be small...', Sec. III D). The benchmark in Sec. IV B covers only GHZ systems with n≤4 and T≤80; no truncation-error estimate or k_cut convergence study is reported for H6. If the conjecture fails, the reported gap values in Fig. 5 and Fig. 6 could be truncation artifacts. Please provide a convergence check for H6 at representative R, or a rigorous bound on the discarded-state contribution.","section":"Sec. III D and Sec. IV B/C"},{"comment":"There is a boundary-condition inconsistency between the matrix in Eq. (15), which is the open-boundary path Laplacian, and the eigenvalues in Eq. (22), which are the periodic eigenvalues 1−cos(2πk/(T+1)). The exact gap of Eq. (15) is 1−cos(π/(T+1)) ≈ π²/(2(T+1)²), not 2π²/(T+1)². More importantly, the periodic momentum states of Eq. (33) do not diagonalize the open-chain H_kin; the block-diagonal form used in Eq. (34) therefore assumes a different, periodic clock Hamiltonian. The Θ(D_c^{-2}) scaling survives, but the numerical Hamiltonian and its off-diagonal momentum couplings need to be corrected or the periodic convention made explicit with the corresponding wrap-around term.","section":"Sec. III B, Eqs. (15), (22), (33), (34)"}],"minor_comments":[{"comment":"The Hamiltonian definition is inconsistent between the body and the appendix: Eq. (38) and Eq. (34) use s[(I_clock−|+><+|)⊗I_sys + I_clock⊗H_target], while Eq. (A1) writes s(|+><+|⊗I_sys + I_clock⊗H_target) and Eq. (A11) omits the δ_kj s I_sys term. Please align the notation.","section":"Appendix A"},{"comment":"There are several typos: 'sepctral' (Sec. III B), 'conjuncture' (Sec. III D), 'drwabacks' (Sec. III C), and 'cv detailed derivation' (Sec. III D). The phrase 'we conjuncture' should be 'we conjecture.'","section":"Throughout"},{"comment":"The caption lists g=0 for the upper curve, but the MPS parameter family is defined for g∈[-1,0), and g=0 is approached as a limit. Please write g→0^{-}.","section":"Fig. 3"},{"comment":"If the authors intend the open-boundary matrix of Eq. (15), the eigenvalues should be 1−cos(πk/(T+1)), k=0,...,T; if they intend periodic boundary conditions, Eq. (15) must include a term coupling |T><0| and |0><T|. The current text mixes both conventions.","section":"Eq. (22)"}],"recommendation":"major_revision","confidential_remarks":"The paper has a promising central idea and the gap-scaling derivation is largely sound, but the numerical headline claim needs to be re-evaluated in absolute terms, and the truncation conjecture must be supported by convergence evidence. The boundary-condition inconsistency is fixable but should not be left as is. I would be willing to review a revised version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a genuinely useful framework paper, but the advertised two-fold gap improvement for H6 is provisional because the momentum-space truncation that produces it is explicitly conjectured and not benchmarked at that system size.\n\nThe genuinely new part is packaging the Feynman–Kitaev clock Hamiltonian as a universal parent Hamiltonian for adiabatic warm starts. Any circuit-preparable state can become the ground state of an initial Hamiltonian, with a clean idling/overlap tradeoff and a Θ(D_c^{-2}) initial gap for depth D_c. Extending to probabilistic circuits via amplitude amplification and getting Θ(p_success/D_c^2) is a nice, non-obvious touch. The rotating-frame momentum-space decomposition is also a practical contribution that could be reused for classical simulations of clock-augmented dynamics—provided the truncation is justified.\n\nThe soft spots are real. Section III D explicitly conjectures that the off-diagonal couplings between slow and fast clock modes are small, citing prefactors O(1/T) and O(D_c/T). The second prefactor is not obviously small in the regime used: the H6 clock is padded to 0.86 overlap, which for typical intermediate overlaps allows D_c/T ~ O(1). The truncation benchmark is limited to GHZ with n≤4 and T≤80, and no truncation error is reported for H6. So the factor-of-two in Fig. 5 could be a truncation artifact. The authors should either prove the conjecture (or a weaker version) or provide explicit convergence checks for H6. Also, the eigenvalues in Eq. (22) are for a periodic Laplacian while the matrix in Eq. (15) is open-boundary; the Θ(T^{-2}) scaling is unaffected, but the exact ω_k is wrong. The claim that the initialization penalty leaves the scaling unchanged is asserted without proof; this is probably fixable but should be shown.\n\nNone of this kills the central idea. The theoretical framework is self-contained, the gap scaling for the clock is standard, and the paper is upfront about its unproven step—which is more than many papers do. The protocol is not circular: the demonstrations use warm starts in the target phase but don't fit any free parameters.\n\nWho's this for? People working on adiabatic state preparation, quantum chemistry initial states, and parent Hamiltonian constructions. It deserves a serious referee, though I would expect heavy revision on the numerical side. I'd bring it to reading group and would cite the framework, with the caveat about the numerics.","headline":"Worth engaging: a clean framework for warm-starting ASP via clock Hamiltonians, but the headline H6 gap improvement rests on an unverified truncation conjecture.","tokens_in":19084,"tokens_out":3367,"would_cite":true,"duration_ms":29650,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Any quantum state with a known preparation circuit can be made the starting point of adiabatic ground-state preparation via the Feynman–Kitaev clock Hamiltonian as a universal parent Hamiltonian.","keywords":["adiabatic state preparation","clock Hamiltonian","parent Hamiltonian","spectral gap","matrix product state","quantum phase transition","quantum chemistry","warm start"],"falsifier":"Compute the full adiabatic gap without momentum truncation for the GHZ-target system at a size where the full basis is still diagonalizable (e.g., n=5 with T=80) and compare with the truncated gap at the same retained-mode fraction; if the relative error exceeds 10% or grows with n rather than decaying with the retained fraction, the truncation conjecture is false.","tokens_in":18171,"feed_emoji":"⚛️","tokens_out":5204,"duration_ms":48258,"temperature":0.7,"pith_summary":"This paper tries to solve a bottleneck in adiabatic state preparation (ASP): when the starting Hamiltonian's ground state lies in a different quantum phase from the target, the path crosses a first-order phase boundary and the spectral gap shrinks exponentially, making ASP impractical. The authors propose UPHAWS, a protocol that encodes the circuit preparing any desired warm-start state into a Feynman–Kitaev clock Hamiltonian, whose ground state (the history state) contains that state as a component. Because the clock Hamiltonian is universal, it lets ASP begin in the same quantum phase as the target, sidestepping the exponentially small gaps. The paper shows the initial gap scales as Θ(D_c^{-2}) in circuit depth, analyzes probabilistic preparation via amplitude amplification, and benchmarks on a GHZ-target MPS family and the linear H6 chain, where a bond-dimension-4 MPS warm start roughly doubles the minimum gap relative to Hartree–Fock. A momentum-space truncation of the clock register makes the classical benchmark simulations tractable, at the cost of an unproved conjecture that discarded fast clock modes do not affect the gap.","feed_headline":"Clock Hamiltonian makes adiabatic warm starts universal","feed_subtitle":"Encoding a state's preparation circuit as a parent Hamiltonian lets adiabatic paths start inside the target quantum phase.","key_machinery":"The central object is the Feynman–Kitaev clock Hamiltonian H_clock = H_init + H_prop, built from the circuit that prepares the warm-start state; its ground state is the history state, a uniform superposition over time slices of the circuit's intermediate states. Padding with identity gates (idling) concentrates probability on the target state. In the momentum basis the propagation term becomes a discrete 1D Laplacian with eigenvalues 1−cos(2πk/(T+1)), giving the Θ(T^{-2}) = Θ(D_c^{-2}) initial gap. The numerical simulations use a rotating-frame, momentum-space truncation that keeps only low-momentum clock modes; its validity rests on an explicit conjecture about small off-diagonal couplings","core_discovery":"Central claim: the Feynman–Kitaev clock Hamiltonian, whose ground state is the history state of a preparation circuit, is a universal parent Hamiltonian for any state with a known circuit. Idling the circuit makes the history state overlap nearly perfectly with |+⟩_clock⊗|ψ_target⟩. The initial gap is Θ(D_c^{-2}) for deterministic depth D_c; amplitude amplification gives Θ(p_success/D_c^2) for probabilistic circuits, and Ω(1/(D^2 D_c^2)) for MPS of bond dimension D. Benchmarks on a GHZ-target MPS family and on linear H6 show the minimum gap stays open when the warm start lies in the target's phase: for H6, a D=4 MPS warm start keeps the rescaled gap above 0.87 versus 0.27 from Hartree–Fock.","pith_inferences":["Editorial inference: the paper's phase-based picture suggests a testable design rule—choose a warm-start ansatz whose order parameter or symmetry sector matches the target phase; the gap along the adiabatic path then serves as a direct diagnostic of phase mismatch.","Editorial inference: if the momentum-truncation conjecture holds, the block-Hamiltonian construction could become a general classical tool for estimating ASP gaps far beyond the small systems simulated, since cost scales with the number of retained slow modes rather than the full clock length.","Editorial inference: the clock register multiplies the system size, so a natural next step—explicitly left open by the paper—is constructing parent Hamiltonians that avoid the ancilla overhead while preserving the warm-start property.","Editorial inference: the authors' framing sharpens the classical–quantum competition: any classical ansatz preparable by a short circuit can be upgraded into a quantum starting point by UPHAWS, provided the phase is known, which reframes the open question of whether efficient quantum state preparation always implies a tractable classical ansatz."],"forward_implications":["If correct, any classical or quantum ansatz with an efficient preparation circuit—MPS, stabilizer states, unitary coupled cluster, or a purely quantum circuit ansatz—can be converted into an ASP starting point, removing the requirement that the initial state be classically tractable.","The Θ(D_c^{-2}) initial gap bound implies the warm-start overhead is polynomial in circuit depth, so the protocol is efficient whenever the preparation circuit is polynomial and the warm-start state lies in the same phase as the target.","The H6 results indicate that a bond-dimension-4 MPS warm start roughly doubles the minimum adiabatic gap relative to Hartree–Fock across the strongly correlated stretched-bond regime, converting classical bond dimension directly into adiabatic gap improvement.","For probabilistic preparation, the gap shrinks only polynomially with bond dimension after amplitude amplification, keeping measurement-based MPS preparation viable as a warm start."],"fun_headline_variants":["Universal parent Hamiltonian boosts adiabatic warm starts","Clock Hamiltonian enables adiabatic warm starts in any phase","MPS warm start doubles energy gap in H6 adiabatic path","Adiabatic warm starts with universal parent Hamiltonians","Warm-start adiabatic paths using Feynman-Kitaev clock"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The paper's numerical evidence rests on an explicitly conjectured assumption: that off-diagonal couplings between slow and fast clock modes are small compared with the O(1) fast-mode energies, so that truncating the clock momentum basis to low modes does not change the spectral gap.","fun_headline_variants_meta":{"raw":{"variants":["Universal parent Hamiltonian boosts adiabatic warm starts","Clock Hamiltonian enables adiabatic warm starts in any phase","MPS warm start doubles energy gap in H6 adiabatic path","Adiabatic warm starts with universal parent Hamiltonians","Warm-start adiabatic paths using Feynman-Kitaev clock"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000298,"raw_usage":{"total_tokens":1620,"prompt_tokens":863,"completion_tokens":757,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":607,"completion_tokens_details":{"reasoning_tokens":676}},"tokens_in":607,"tokens_out":757,"duration_ms":7500,"temperature":1.0,"reasoning_tokens":676,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T19:43:34.200201+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the full adiabatic gap without momentum truncation for the GHZ-target system at a size where the full basis is still diagonalizable (e.g., n=5 with T=80) and compare with the truncated gap at the same retained-mode fraction; if the relative error exceeds 10% or grows with n rather than decaying with the retained fraction, the truncation conjecture is false.","supporting_citations":[],"review_version":1}