{"id":"174754af-019f-4d63-99a9-b8278ec4083b","arxiv_id":"2608.11369","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":4,"one_line_summary":"A measurement-based holonomic scheme is constructed for cat and GKP bosonic codes, with the target gate set as the endpoint of the control path.","lead":"This paper proposes using continuous measurements, instead of direct Hamiltonian driving, to perform logical quantum gates on two types of bosonic error-correcting code, cat and GKP. It gives explicit paths and formulas for Z rotations on cat codes and a T gate on GKP codes, plus leakage estimates.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"GKP instantaneous projector in Eq (57) is not a projector: S_Z(t) is non-Hermitian and S_X,S_Z are not involutions, so the GKP holonomy and Zeno argument lack a well-defined measured object.","rationale":"The reader's weakest assumption correctly identifies the GKP projector in Eq. (57) as the central problem. The operator is not a projector: S_Z(t) is a non-Hermitian displacement multiplied by a phase, and S_X(t) is a translation with a state-dependent phase; neither has eigenvalues restricted to ±1, so (I+S)/2 is not an orthogonal projection. An explicit position-basis check at t=0 shows P_GKP fails to annihilate |√π>, which is outside the GKP code. Since the measured observable must be a projector for the Zeno argument and for the Grassmannian holonomy framework of Appendix A, the GKP section's derivation of the T_GKP gate is unsupported as written. The cat-code half appears internally consistent and does not suffer from this flaw. The endpoint issue (V(tf) set equal to the target gate) is gauge-dependent and less decisive, because the physical operation V(tf)h(tf) is gauge-invariant; the invalid projector is the concrete, load-bearing mathematical gap. The finite-energy projector P_ϵ in Eq. (70) could provide a route to repair the GKP construction, but that repair is not made in the paper. The reader's CONDITIONAL verdict therefore stands without adjustment: the paper is plausible for cats but needs a corrected GKP projection and a re-derived holonomy before the GKP claim can be accepted.","tokens_in":20394,"tokens_out":19619,"duration_ms":180538,"concrete_test":"Evaluate Eq. (57) at t=0 on the position eigenstate |q=√π> (or on a finite-energy regularization). If (I+S_Z)/2 leaves the state unchanged while (I+S_X)/2 maps it to (|√π>+|3√π>)/2, the operator does not annihilate the orthogonal complement of the GKP code, proving it is not the code projector. Equivalently, check P_GKP(0)^2≠P_GKP(0) and P_GKP(0)†≠P_GKP(0) in a truncated Fock basis. If confirmed, recompute the GKP holonomy with a valid instantaneous projector (e.g., V(t)P_ϵ V† with ϵ→0) to see whether T_GKP still emerges.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing flaw is Eq. (57). It defines P_GKP(t)=1/4(I+S_X(t))(I+S_Z(t)) as the instantaneous code projector. But S_Z(t)=e^{i2πg(t)}S_Z (Eq. 55) is unitary, not Hermitian, and S_X(t) (Eq. 56) is a displaced/phase-twisted translation; neither operator squares to I on the oscillator Hilbert space. (I+S)/2 is a spectral projection only when S=S† and S^2=I. Consequently P_GKP(t) is not Hermitian and not idempotent. Even at t=0, P_GKP|√π>_Q = 1/2(|√π>+|3√π>)_Q, which is not in the ideal GKP code, so Eq. (57) does not project onto span{|0_L>,|1_L>}. The subsequent Wilczek-Zee connection, horizontal lift, dressed KL conditions Eq. (62), and leakage bound are therefore derived for an object that is not the monitored code subspace; continuous measurement of a non-Hermitian non-idempotent operator is not a valid Zeno protocol. The finite-energy P_ϵ in Eq. (70) is a legitimate projector but is used only for leakage, not for the holonomy path. This invalidates the GKP T_GKP claim as stated.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper applies continuous-measurement-based holonomic quantum computation (CMHQC) to two bosonic encodings. For the four-component cat code, it constructs a squeezed-cat path V(t)=e^{iφ(t)Z_L}S(ξ(t)) with V(0)=I and V(t_f)=e^{iθZ_L}, chooses the squeezing amplitude through the Bessel zero j_{0,1} to enforce the horizontal condition, and claims arbitrary logical Z rotations including the non-Clifford T_L gate. For the GKP code, it constructs V(t)=e^{iκ(t)f(Q/√π)}e^{-i√π g(t)P} with V(t_f)=T_GKP, introduces an instantaneous projector P_GKP(t)=(I+S_X(t))(I+S_Z(t))/4, and claims a purely geometric implementation of the logical T_GKP gate. The paper also derives dressed Knill-Laflamme conditions and analytic leakage bounds for both codes, with numerical fidelities and leakage simulations.","tokens_in":20768,"tokens_out":15429,"duration_ms":152973,"significance":"If the constructions were valid, the paper would be a valuable step toward measurement-based holonomic gates in bosonic error-correcting platforms, especially because it targets non-Clifford gates and includes explicit error-correction and leakage analysis. The cat-code half is built from legitimate projectors, contains an analytic horizontal-lift calculation, and is supported by leakage simulations; that part is a defensible contribution. The GKP half, however, rests on an invalid definition of the instantaneous code projector, so the central GKP claim is not supported as written. The paper is also careful to state approximations and to provide numerical checks, which is a strength, but the GKP numerical fidelity only tests the endpoint unitary V(t_f), not the measurement-induced Zeno dynamics.","major_comments":[{"comment":"The operator P_GKP(t)=1/4(I+S_X(t))(I+S_Z(t)) is not a projector on the oscillator Hilbert space. From Eq. (55), S_Z(t)=e^{i2πg(t)}S_Z is unitary but not Hermitian, and S_X(t) in Eq. (56) is a displaced translation; neither satisfies S=S^† or S^2=I. Therefore (I+S)/2 is not a spectral projection, and P_GKP(t) is neither Hermitian nor idempotent. As a concrete check at t=0, P_GKP|√π>_Q = 1/2(|√π>_Q+|-√π>_Q), which is not an element of the ideal GKP code. Consequently Eq. (57) does not project onto span{|0_L>,|1_L>}, and the subsequent Wilczek-Zee connection, horizontal-lift condition, dressed Knill-Laflamme conditions in Eq. (62), and Zeno leakage argument are derived for an object that is not the monitored code subspace. The finite-energy projector P_ϵ in Eq. (70) is legitimate but is used only for the leakage estimate, not for the holonomy path. This invalidates the GKP T_GKP claim as stated.","section":"Section III.B, Eq. (57)"},{"comment":"The projected cubic-phase calculation is not well-defined in the ideal-GKP limit. Equations (D5) and (D6) express the matrix elements of f(Q/√π+g) as infinite sums over the ideal comb states, such as ∑_j f(2j√π), but for the cubic polynomial f in Eq. (48) these sums diverge. The step of 'ignoring the modulo 2 values' before Eq. (D7) is not a controlled regularization, so the resulting formulas a_I(t) and a_Z(t), the integral condition Eq. (D11), and the conclusion that h(t_f) is a trivial phase are formal rather than mathematically justified. A consistent treatment must either use a finite-energy projector such as P_ϵ throughout or define the projected operators only through their action on logical basis states with an explicit regularization.","section":"Appendix D.1, Eqs. (D5)-(D7)"},{"comment":"The claim that the logical gates arise from the projected Wilczek-Zee connection is not supported by the equations as written. In both protocols the target gate is imposed as the endpoint of the control unitary, V(t_f)=exp(iθZ_L) for the cat code and V(t_f)=T_GKP for the GKP code, and the horizontal condition is then used to make the auxiliary factor h(t_f) a trivial phase (Eqs. (C15) and (D11)-(D14)). By Eq. (A16b), the final state is therefore |ψ(t_f)>=V(t_f)|ψ(0)> up to a global phase, so the logical operation is exactly the endpoint of the control unitary. This can be a legitimate holonomic protocol if V(t_f) is interpreted as the holonomy of the closed Grassmannian path, but the paper should state this identification explicitly; as written, the abstract's statement that the connection 'generates' the logical operations is contradicted by the fact that the projected connection is made to vanish.","section":"Eqs. (21b), (51b), and (A16b)"}],"minor_comments":[{"comment":"The normalization constant N_k is typeset in a way that is difficult to parse; the parentheses and exponentials should be rewritten so that the large-|α| limit N_k→1/2 is immediately visible.","section":"Eq. (3)"},{"comment":"The definition of p_code as p_0(t_f)+p_1(t_f) needs clarification: for a superposition initial state the probability of remaining in the code is <ψ|P_0|ψ>, and the derivation appears to treat the two logical branches as if they were incoherent alternatives.","section":"Section II.D, Eqs. (C41)-(C42)"},{"comment":"The fidelities plotted are overlaps with V(t_f)|ψ(0)>, not fidelities of the full measurement-based Zeno evolution; the captions should say this explicitly to avoid overstating the numerical verification of the CMHQC protocol.","section":"Figs. 1 and 4"},{"comment":"The displayed simplification of S_X(t) is correct only up to the stated replacement f(x)-f(x-2); the intermediate expression should be written out to avoid confusing the reader about the order of factors in the exponential.","section":"Eq. (56)"}],"recommendation":"major_revision","confidential_remarks":"The GKP section has a load-bearing flaw: Eq. (57) does not define a projector, and the formal ideal-GKP matrix elements in Appendix D.1 require regularization. This cannot be fixed by a small edit; the authors would need to redo the GKP horizontal-lift and Knill-Laflamme derivations with a genuine instantaneous projector, most likely the finite-energy projector P_ϵ, or substantially weaken the GKP claims. If that rework is not possible, the paper should be restricted to the cat-code results. The cat-code construction appears sound within its stated large-|α| and strong-measurement approximations, but the authors should also clarify the sense in which the endpoint unitary V(t_f) constitutes a holonomic gate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the cat-code section is a legitimate extension of your earlier CMHQC work and the calculations hold together in the large-|alpha| limit. The GKP section does not: the object in Eq. (57) is not a projector, so the claim that the T gate is a measurement-induced holonomy is unsupported as written. The same endpoint-imposition problem appears in the cat half, though less damagingly.\n\nWhat is new: applying CMHQC to bosonic codes at all, the squeezed-cat trajectory, the translated-lattice path, and the dressed Knill-Laflamme/leakage analysis. The cat derivation is careful: the projected connection, the Bessel-zero choice for the horizontal lift, and the leakage bound all work within the stated approximations. The finite-energy GKP phase-error calculation in Appendix D is also useful and matches the Strawberry Fields numerics.\n\nSoft spots. First, Eq. (57) defines P_GKP(t) as (I+S_X(t))(I+S_Z(t))/4. S_Z(t)=e^{i2pi g(t)}S_Z is unitary but not Hermitian, and S_X, S_Z are displacement operators, not involutions. This is not a projector, so the Zeno argument, the dressed KL conditions, and the leakage bound for the instantaneous GKP space have no well-defined measured subspace. At t=0 the formula sends |sqrt(pi)> to (|sqrt(pi)>+|3sqrt(pi)>)/2, which is not the action of a rank-2 projection. The finite-energy P_epsilon in Eq. (70) is a genuine projector, but it is used only for leakage, not for the holonomy path. Second, both halves set V(tf) equal to the target gate (Eqs. (21b) and (51b)) and then choose parameters so the extra holonomy h is trivial. If h is trivial, the gate is just the endpoint of the control unitary V(tf), not something generated by the geometric connection. The phrase \"without Hamiltonian control\" is misleading: V(t) is a unitary control path; the measurement enforces Zeno confinement but does not produce the gate. Minor: the numerical fidelities check V(tf) on the initial state, not the full monitored evolution.\n\nWho this is for: people working on CMHQC or bosonic holonomic control. The cat section deserves a careful read; the GKP section needs a real fix, starting with a correct designation of the instantaneous GKP projector and a clearer statement of what the measurement adds. Send to peer review, but flag the GKP projector and the endpoint-imposition issue as mandatory revisions. This is not a desk reject: one half is solid, the other is currently unsupported.","headline":"Cat-code half is a plausible CMHQC extension; the GKP half has a load-bearing projector error, and both halves build the target gate into the path endpoint rather than deriving it from the holonomy.","tokens_in":21256,"tokens_out":6738,"would_cite":false,"duration_ms":60836,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P68","81P70"],"pacs":["03.67.Lx","03.67.Pp"],"model":"deepseek-v4-flash","headline":"Continuous measurement of a time-dependent code subspace can implement logical T gates in cat and GKP bosonic codes through geometric holonomy alone.","keywords":["holonomic quantum computation","continuous measurement","Zeno effect","cat code","GKP code","Wilczek-Zee connection","non-Clifford T gate","Knill-Laflamme conditions"],"falsifier":"For the GKP claim, evaluate $P_{\\mathrm{GKP}}(t)^2-P_{\\mathrm{GKP}}(t)$ with $P_{\\mathrm{GKP}}(t)=\\tfrac14(I+S_X(t))(I+S_Z(t))$ at an intermediate $g(t)\\neq 0$; if this operator does not vanish, then Eq. (57) is not a projector and the dressed Knill-Laflamme proof collapses. For the cat claim, simulate the actual Lindblad measurement dynamics at finite $\\kappa$ with $|\\alpha|\\approx 3$ and check whether the recovered logical operation has a residual $Z_L$ or $Y_L$ error beyond $O(e^{-|\\alpha|^2})$; if the residual does not shrink as $|\\alpha|$ grows, the trajectory does not implement $\\exp(i\\theta Z_L)$.","tokens_in":20194,"feed_emoji":"⚛️","tokens_out":9569,"duration_ms":84187,"temperature":0.7,"pith_summary":"Continuous measurement, not Hamiltonian driving, can be the engine of non-Clifford logical gates in bosonic error-correcting codes. The paper constructs two explicit closed paths on the Grassmannian manifold of code subspaces: a squeezed-cat path for four-component cat codes that yields arbitrary rotations about the logical Z axis, including the T gate, and a translated-lattice path for GKP codes that yields the logical T_GKP gate as a geometric holonomy. The protocols rely on Zeno confinement to keep the state in the instantaneous code space while the projected Wilczek-Zee connection accumulates the logical operation, and the paper verifies dressed Knill-Laflamme conditions and bounds leakage at finite measurement strength. A reader should care because this replaces Hamiltonian engineering of non-Clifford gates with a measurement-induced geometric mechanism on the two leading hardware-relevant bosonic encodings.","feed_headline":"Continuous measurement alone can produce the bosonic T gate","feed_subtitle":"Monitoring a moving code space yields the T gate as a geometric phase, with no Hamiltonian control.","key_machinery":"The central mechanism is Zeno confinement of a time-dependent rank-two projector $P(t)=V(t)P_0 V^\\dagger(t)$, combined with the parallel-transport (horizontal-lift) condition on the Stiefel bundle, $L^\\dagger(t)\\dot L(t)=0$. In the rotating frame, the projected Wilczek-Zee connection acts as an effective Hamiltonian; the protocol chooses the driving functions so that the integrated connection $h(t_f)$ is a trivial phase, leaving the geometric action of $V(t_f)$ on the code space as the logical operation. For cats the path is generated by two-photon squeezing plus a logical Z phase; for GKP it is generated by a cubic-phase polynomial $f(Q/\\sqrt{\\pi})$ and a momentum translation $g(t)$, with $A$ selected to cancel the accumulated $Z_L$ term. The error-correction analysis works with dressed errors $\\tilde E_i(t)=V^\\dagger(t)E_i V(t)$ and shows that the projected sandwich $P(t)E_j^\\dagger E_i P(t)$ has the same vanishing/Pauli block structure as the ideal code.","core_discovery":"The paper claims that continuous measurement of a time-varying code subspace alone, without any logical Hamiltonian term, can implement holonomic gates in bosonic codes. For the four-component cat code, the unitary path $V(t)=e^{i\\varphi(t)Z_L}S(\\xi(t))$ with $\\varphi(t)=\\theta t/t_f$ and squeezing amplitude $r(t)=j_{0,1}(2|\\alpha|^2)^{-1}\\sin(2\\pi t/t_f)$ closes the loop and satisfies $V(t_f)=e^{i\\theta Z_L}$; choosing $\\theta=\\pi/8$ gives the non-Clifford $T_L$ gate, with numerical process infidelity below $10^{-3}$ for $|\\alpha|\\gtrsim 3$. For the GKP code, the path $V(t)=e^{i\\kappa(t)f(Q/\\sqrt{\\pi})}e^{-i\\sqrt{\\pi}g(t)P}$, with $\\kappa(t)=2\\pi t/t_f$, $g(t)=A\\sin^2(\\pi t/t_f)$, and $A=-(8\\pm 2\\sqrt{7})/9$ chosen so the integrated $Z_L$ component of the projected connection vanishes, yields $V(t_f)=e^{2\\pi i f(Q/\\sqrt{\\pi})}=T_{\\mathrm{GKP}}$. The paper further claims that the instantaneous code spaces satisfy dressed Knill-Laflamme conditions for the relevant error models, and that leakage out of the code space is bounded by analytically derived Zeno estimates that vanish in the strong-measurement limit.","pith_inferences":["The GKP construction's cubic polynomial $f$ is not essential: any phase-space function that acts diagonally on the GKP lattice and whose integrated projected $Z_L$ term can be cancelled should generate a different diagonal logical gate from the same translated-lattice template, potentially covering additional non-Clifford rotations.","Because the paper's finite-energy analysis isolates the dominant effect as a renormalized logical phase rather than state mixing, a natural testable extension is to run the protocol with finite-energy GKP states and verify that the gate error tracks the predicted $\\delta\\theta(\\epsilon)$ rather than an additional leakage channel.","If the same horizontal-lift construction is combined with a two-qubit geometric gate, the two single-mode protocols would yield a measurement-only universal gate set for bosonic hardware; the paper itself does not construct that two-qubit gate.","An experimental signature of the cat protocol is the predicted plateau in code-space population as $\\kappa t_f$ grows, with the leakage scaling fixed by $j_{0,1}^2\\pi^2/(\\kappa t_f |\\alpha|^2)$; measuring this scaling in a cavity-QED setup would directly test whether the Zeno limit really implements the geometric gate."],"forward_implications":["Choosing $\\theta=\\pi/8$ in the cat trajectory yields a non-Clifford $T_L$ gate whose numerically computed process infidelity falls below $10^{-3}$ once $|\\alpha|\\gtrsim 3$.","The GKP translated-lattice path realizes $T_{\\mathrm{GKP}}$ without a cubic-phase Hamiltonian, so the non-Clifford gate is fixed by the geometry of the closed measurement path rather than by a directly engineered logical interaction.","Because the instantaneous code spaces obey dressed Knill-Laflamme conditions for the relevant bosonic error models, error correction can in principle operate while the holonomic evolution is underway.","Leakage from the code space is suppressed by strong measurement: the cat bound scales as $1-\\exp\\big(-(8/\\kappa)\\int_0^{t_f}\\|B(t)\\|_2^2\\,dt\\big)$ and the GKP bound as $O(1/(\\kappa t_f))$, so increasing the Zeno parameter drives both protocols toward the ideal geometric limit."],"supporting_citations":[{"why":"defines the CMHQC framework of continuously monitored Grassmannian paths and Zeno confinement that this paper extends to bosonic codes.","marker":"[1]"},{"why":"supplies the steering and parallel-transport formalism that the paper adapts for cat and GKP trajectories.","marker":"[2]"},{"why":"introduces the Wilczek-Zee connection whose holonomy is the working mechanism of the gates.","marker":"[3]"},{"why":"establishes holonomic quantum computation as a mechanism for logical gates, the concept applied here.","marker":"[4]"},{"why":"defines the GKP code, its stabilizers, and the cubic-phase $T_{\\mathrm{GKP}}$ gate that the GKP protocol aims to realize.","marker":"[6]"},{"why":"provides the four-component cat-code manifold and the engineered two-photon-loss setting used by the cat protocol.","marker":"[8]"}],"fun_headline_variants":["Continuous measurement alone implements bosonic T gates","No Hamiltonian control: bosonic T gate from measurement","Geometric T gate from Zeno measurement in bosonic codes","Measurement-only holonomic gates for GKP and cat codes","Watch a code space move to get the T gate"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing assumption is that the Zeno limit faithfully realizes the idealized parallel-transport evolution while all neglected corrections, namely $O(e^{-|\\alpha|^2})$ terms for the cat path and the non-Hermiticity of the translated GKP stabilizer $S_Z(t)=e^{i2\\pi g(t)}S_Z$, remain negligible over the whole trajectory.","fun_headline_variants_meta":{"raw":{"variants":["Continuous measurement alone implements bosonic T gates","No Hamiltonian control: bosonic T gate from measurement","Geometric T gate from Zeno measurement in bosonic codes","Measurement-only holonomic gates for GKP and cat codes","Watch a code space move to get the T gate"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000652,"raw_usage":{"total_tokens":3062,"prompt_tokens":1087,"completion_tokens":1975,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":703,"completion_tokens_details":{"reasoning_tokens":1898}},"tokens_in":703,"tokens_out":1975,"duration_ms":13681,"temperature":1.0,"reasoning_tokens":1898,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:14:52.091844+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For the GKP claim, evaluate $P_{\\mathrm{GKP}}(t)^2-P_{\\mathrm{GKP}}(t)$ with $P_{\\mathrm{GKP}}(t)=\\tfrac14(I+S_X(t))(I+S_Z(t))$ at an intermediate $g(t)\\neq 0$; if this operator does not vanish, then Eq. (57) is not a projector and the dressed Knill-Laflamme proof collapses. For the cat claim, simulate the actual Lindblad measurement dynamics at finite $\\kappa$ with $|\\alpha|\\approx 3$ and check whether the recovered logical operation has a residual $Z_L$ or $Y_L$ error beyond $O(e^{-|\\alpha|^2})$; if the residual does not shrink as $|\\alpha|$ grows, the trajectory does not implement $\\exp(i\\theta Z_L)$.","supporting_citations":[{"cited_title":"Continuous measurement-based holonomic quantum computation","cited_arxiv_id":"2510.06725","evidence_quote":"defines the CMHQC framework of continuously monitored Grassmannian paths and Zeno confinement that this paper extends to bosonic codes."},{"cited_title":"Lanka, J","cited_arxiv_id":null,"evidence_quote":"supplies the steering and parallel-transport formalism that the paper adapts for cat and GKP trajectories."}],"review_version":1}