{"id":"4e6ba490-af00-4ef0-9b06-601e2107f4f9","arxiv_id":"2607.27144","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"Parity (even/odd) time-bin encoding makes a fixed delay a bidirectional logical flip, enabling a three-CNOT optical SWAP between polarization and time-bin qubits on one photon.","lead":"The paper defines a parity-based time-bin qubit (even vs odd arrival slots) so a fixed delay flips the logical bit both ways. That encoding lets three standard optical CNOTs form a deterministic SWAP between a photon’s polarization and its time-bin state.","discovery_kind":"new_method","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The unitary-level SWAP construction is correct; the load-bearing risk is that timing jitter and finite wavepacket width make parity itself ill-defined at the EOM sampling instant, and the paper's own flip-error formula (Eq. 13) is inverted, so the voltage-drift budget is misstated.","rationale":"This is a proposal paper, and its core is sound: the parity encoding genuinely converts a fixed delay into a bidirectional logical flip, which conventional early/late encoding cannot do, and I could not find a flaw in the circuit identity or the component-to-gate mapping — both check out by hand, including on entangled/superposition inputs. The honest load-bearing concern is realizability of clean parity labeling through the cascade, which the reader already identified; my pass sharpens it in two ways. First, the failure mechanism is specific: parity becomes ill-defined when τ_p + k·σ_jitter ≳ Δt at the EOM sample instant, and that error is coherent and uncorrectable, unlike loss. Second, the paper's one quantitative fidelity result (Eq. 13) is inverted — a concrete, checkable internal error in a section the abstract advertises as a contribution — which, together with the 1σ jitter treatment in Eq. 14 and the Discussion's disavowal of the very hardware analysis Sec. V attempts, means the error budget does not currently support the stated Δt = 1 ns / 400 MHz design point. This warrants fixing but not rejection: the central architectural claim stands conditionally. A secondary limitation worth noting (not load-bearing here): the SWAP output inherits input-dependent absolute time shifts (0–2Δt), so composing this primitive into circuits requiring two-photon interference would need re-synchronization; within the paper's stated single-photon routing scope that is acceptable. Recommended verdict: UNCHANGED (CONDITIONAL, MODERATE).","tokens_in":8575,"tokens_out":8831,"duration_ms":54696,"concrete_test":"Monte-Carlo or analytic convolution: model the photon as a Gaussian wavepacket (τ_p = 50–100 ps) arriving with Gaussian jitter (σ = 50 ps) at an EOM driven by a trapezoidal waveform (τ_switch = 100 ps, period 2Δt, flat-top V_π). Integrate the fraction of the photon's temporal probability mass sampling the wrong voltage level (parity misassignment) times the corrected flip error sin²(πε_V/2) with ε_V = 1%, versus Δt. The Sec. V design point (Δt = 1 ns) should yield total per-stage error < 10⁻³ at a ≥5σ confidence margin, not the 1σ allowance the paper uses; if it does not, the cited Δt and the ~400 MHz rate claim fail and the encoding premise needs a larger Δt.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I verified the ideal claim directly: Eqs. 4–8 are the standard three-CNOT identity, and the physical maps (Eqs. 9–10) implement the correct CNOTs. Tracing superposition inputs (e.g., (α|H⟩+β|V⟩)|E⟩ and |H⟩(γ|E⟩+δ|O⟩)) through the actual three-stage hardware reproduces the SWAP, with input-dependent absolute bin shifts of 0, 1, or 2Δt harmlessly absorbed by the parity label. So the weakest point is not the logic but the premise that parity is a well-defined binary label at every stage. The entire encoding collapses if the photon's temporal wavepacket plus cumulative jitter straddles a bin boundary at the EOM sampling instant: then CNOT_{T→P} applies the wrong rotation coherently, and the error is uncorrectable because which-bin information is the logical information. Two signs the paper's analysis understates this: (1) Eq. 13 is inverted — with V = V_π(1+ε), P_flip = sin²(π(1+ε)/2) = cos²(πε/2) is the success probability, so the flip error is sin²(πε/2) ≈ (πε/2)², not cos²(πε/2) as printed; as written, Eq. 13 assigns unit error to a perfect drive. (2) Eq. 14 folds σ_jitter in as a 1σ allowance \"rather than a specified statistical confidence margin,\" but a bin-misassignment-limited encoding needs roughly a 5σ margin; the paper's numbers (550 ps budget vs 1 ns Δt) do not clearly provide it. This is the reader's weakest assumption, sharpened to the parity-straddling failure mode with a concrete internal error as evidence.","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript introduces a parity-based time-bin encoding in which logical |0⟩_T and |1⟩_T are even and odd multiples of a spacing Δt, so that a physical delay of Δt implements a bidirectional logical flip — something a conventional early/late encoding cannot do. This encoding makes a PBS-delay-PBS stage implement CNOT_{P→T} and a synchronized EOM (or Kerr switch) implement CNOT_{T→P}; composing three CNOTs via the standard identity yields a deterministic SWAP between the polarization and time-bin qubits of a single photon. Section III.A verifies the SWAP by explicit basis-state expansion (Eqs. 6–8), Section IV maps each CNOT to standard optics (Eqs. 9–10), and Section V gives a Jones-calculus polarization-flip error analysis (Eqs. 11–13) and a timing budget for choosing Δt (Eq. 14) with a nominal 1 ns design point and ~400 MHz ideal gate rate.","tokens_in":9070,"tokens_out":3332,"duration_ms":134379,"significance":"If the encoding holds up experimentally, the paper provides a genuinely useful primitive: to my knowledge the parity (even/odd) time-bin labeling is new, it cleanly removes the one-directionality obstruction of early/late encodings, and the resulting SWAP uses only standard components (PBS, delay line, EOM). The unitary-level claim is fully verified by direct basis expansion rather than asserted, and the paper supplies falsifiable engineering budgets (Eqs. 12–14, the 550 ps vs 1 ns design point) that an experimentalist can test immediately. I checked the central construction independently: Eqs. 4–8 are the correct three-CNOT identity, the optical maps in Eqs. 9–10 implement the correct CNOTs, and tracing superposition inputs through the three-stage cascade reproduces the SWAP, with input-dependent absolute bin shifts of 0–2Δt absorbed harmlessly by the parity label. The ideal-level result is sound; the concerns below are confined to the error/timing analysis and to gaps in the physical model of the encoding, not to the logical construction.","major_comments":[{"comment":"Eq. (13) is inverted. With V = V_π(1+ε_V), the retardance is θ = π(1+ε_V), so Eq. (12) gives P_flip = sin²(π(1+ε_V)/2) = cos²(πε_V/2). That is the *success* probability of the flip; the flip *error* is 1 − P_flip = sin²(πε_V/2) ≈ (πε_V/2)². As printed, Eq. (13) assigns unit error to a perfect drive (ε_V = 0) and zero error to a drive that does nothing (ε_V = 1), so the 'direct specification on allowed drive-voltage drift' claimed in the text is misstated. The fix is a one-line correction (P_error = sin²(πε_V/2)), but because this is the paper's only quantitative rotation-accuracy result, it should be corrected and the implied tolerance (e.g., ε_V ≲ 2% for 10⁻³ error) stated explicitly.","section":"§V.B, Eq. (13)"},{"comment":"The timing analysis treats the photon as a point arrival at a bin center and folds jitter in as a 1σ 'representative timing allowance' (the text under Eq. 14 says this explicitly). For a parity encoding this understates the dominant failure mode: if the photon's temporal wavepacket plus cumulative jitter straddles a bin boundary at the EOM sampling instant, CNOT_{T→P} applies the wrong rotation coherently, and the error is uncorrectable because which-bin information is the logical information. A bin-misassignment-limited encoding needs a confidence-margin analysis (roughly 5σ, or an explicit target misassignment probability), not a 1σ allowance; with σ_jitter = 50 ps the stated 450 ps margin corresponds to ~9σ and is probably adequate, but this should be argued rather than left implicit. The budget should also account for (i) jitter accumulating across the three sequential stages and (ii","section":"§V.A and §V.C, Eq. (14)"},{"comment":"The paper never specifies how the parity qubit is read out. Distinguishing |even⟩ from |odd⟩ requires measuring arrival time modulo 2Δt against an absolute reference t0, with detector resolution well below Δt — a stronger and differently structured requirement than resolving two adjacent early/late bins. Since the whole architecture rests on parity being a well-defined, measurable binary label at every stage, a short subsection on parity readout (and on the physically finite extent of the 'infinite periodic grid', i.e., how many parity bins a real source/detector chain supports) is needed to make the encoding operationally complete.","section":"§II.B and §V.C"}],"minor_comments":[{"comment":"The first sentence ('Multi-degree-of-freedom photonic quantum processing requires routing...') is duplicated verbatim.","section":"Abstract"},{"comment":"The Discussion states the scheme 'does not yet include a quantitative hardware-level analysis for (i) finite EOM rotation accuracy versus drive-voltage error... and (ii) detector/electronics timing resolution and jitter limitations' — but §V.B and §V.C present exactly these analyses. This sentence appears to be a leftover from an earlier draft and should be reconciled with §V.","section":"§VI"},{"comment":"Ref. [15] (Knill, Laflamme, Milburn) appears in the bibliography but is never cited in the text.","section":"References"},{"comment":"The basis-state proof uses a product-state input. This is sufficient by linearity, but one sentence saying so (and noting the SWAP therefore also acts correctly on entangled inputs, e.g., when one DOF is entangled with an external system) would preempt a common reader question, especially given Ref. [6]'s entanglement-transfer use case.","section":"§III.A, Eqs. (6)–(7)"},{"comment":"Only the |H⟩ rows of the EOM action are shown. Including the |V⟩ rows (|odd⟩|V⟩ → |odd⟩|H⟩, |even⟩|V⟩ unchanged) would make the CNOT action explicit and match Eq. (5).","section":"§IV.A, Eq. (9)"},{"comment":"Sentence fragment: 'Equivalently, a timing budget can be written as Δt ≳ τ_open + τ_switch, so each interval includes both transition time and a stable sampling region. where τ_switch is...' — the clause beginning 'where' should be merged with the preceding sentence.","section":"§V.A"},{"comment":"The timing diagrams are schematic and legible, but Fig. 2 would benefit from marking the stable sample points explicitly on the voltage axis (the legend mentions them without showing them), and Fig. 7's 'eye' should indicate the rise/fall time τ_switch so the reader can connect it to Eq. (14).","section":"Figures 1, 2, 7"}],"recommendation":"minor_revision","confidential_remarks":"The manuscript is short and the core idea is simple, but the encoding appears genuinely new and the construction is verified rather than asserted; it reads naturally as a Letter. Ref. [7] is prior work by the second author and is cited appropriately as the conventional-encoding counterpart, not over-claimed. All three major comments are fixable within the existing scope (one is a one-line sign correction); none touches the central unitary-level claim, which I verified independently."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The useful bit here is the encoding, not the SWAP identity. Even/odd parity of a Δt grid turns a one-way physical delay into a logical X on the time-bin qubit, so a PBS-delay stage is a real CNOT_P→T and a clocked EOM half-wave is CNOT_T→P. Three of those give a deterministic polarization–time-bin SWAP on one photon. That is a clean routing primitive for multi-DOF single-photon circuits, and the basis walk-through (Sec. III.A) and optical maps check out at the ideal unitary level.\n\nWhat is actually new is the parity labeling itself. Early/late bins cannot flip late→early with a fixed delay; the paper’s grid can. Prior momentum–polarization SWAP and Kerr conversion work are cited honestly; the three-CNOT decomposition is textbook. The contribution is the design choice that makes ordinary PBS/EOM hardware implement the controlled flips bidirectionally.\n\nSoft spots are real but proportionate. There is no experiment—proposal only. Eq. 13 is inverted: with V = V_π(1+ε) the flip success probability is cos²(πε/2), so the error is sin²(πε/2), not the printed “P_flip error = cos²…”. As written, perfect drive gets unit error. Timing analysis (Eq. 14) folds σ_jitter in as a 1σ allowance; for a parity label that collapses if the wavepacket straddles a bin edge at the EOM sample instant, you want a several-σ margin, and the 550 ps vs 1 ns numbers do not clearly give it. Residual birefringence, path imbalance, and loss asymmetry are mentioned lightly or not at all. Abstract has a copy-paste duplication; Discussion slightly undersells the analysis already in Sec. V. None of that breaks the unitary claim.\n\nWho it is for: people building multi-DOF photonic gates who need a concrete inter-encoding router. Math and citations are solid for a short design note. I would send it to referees; they will demand the error-formula fix, a proper jitter/parity-straddle budget, and either simulation or a path to experiment. Worth engaging if you work in that niche; not a platform-wide result.","headline":"Correct unitary SWAP via a parity time-bin label; real novelty is the encoding, with one inverted error formula and under-analyzed bin-straddling risk.","tokens_in":10105,"tokens_out":558,"would_cite":false,"duration_ms":10858,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Parity-based time-bin encoding lets a single photon swap its polarization and arrival-time qubits using ordinary optics.","keywords":["parity-based time-bin","SWAP gate","polarization qubit","time-bin qubit","electro-optic modulator","multi-degree-of-freedom","CNOT","photonic quantum processing"],"falsifier":"Build the PBS-delay / EOM / PBS-delay cascade at a chosen Δt, inject known polarization–time-bin product states, and test whether the output matches the swapped state within the paper’s predicted flip-error and jitter budgets; systematic mismatch beyond those models would falsify a working physical SWAP.","tokens_in":9795,"feed_emoji":"🔄","tokens_out":875,"duration_ms":35698,"temperature":0.7,"pith_summary":"Photonic circuits that store qubits in several properties of one photon need a clean way to move information between those encodings. Conventional early/late time-bin labels only let a delay push an early photon late; they cannot push a late photon early, so a polarization-controlled delay cannot act as a logical controlled flip. This paper defines logical time-bin states by even versus odd multiples of a fixed spacing Δt. Under that parity rule a physical delay of Δt flips the logical time-bin bit in both directions, a polarization beam-splitter delay line becomes CNOT from polarization to time-bin, and a clocked electro-optic modulator becomes CNOT from time-bin to polarization. Three such CNOTs compose into a deterministic SWAP between the two degrees of freedom, giving a routing primitive for multi-degree-of-freedom single-photon architectures.","feed_headline":"Parity labels make time-bin and polarization qubits swappable","feed_subtitle":"Even/odd arrival times turn ordinary delays and modulators into a deterministic single-photon SWAP.","key_machinery":"Parity-based time-bin encoding: logical |0⟩_T and |1⟩_T are even and odd multiples of Δt, so a delay of exactly Δt implements |0⟩_T ↔ |1⟩_T regardless of where the photon sits inside its bin.","core_discovery":"Defining logical time-bin states as even and odd multiples of a spacing Δt makes a fixed physical delay implement a bidirectional logical bit flip. With that encoding, polarization-controlled delay realizes CNOT from polarization to time-bin, synchronized electro-optic modulation realizes CNOT from time-bin to polarization, and the standard three-CNOT identity therefore yields a deterministic SWAP between polarization and time-bin qubits on a single photon built from ordinary optical components.","pith_inferences":["Loss or phase imbalance between the delayed and undelayed PBS arms is outside the unitary model and would have to be engineered away before the composed SWAP reaches high fidelity.","The same parity grid could support richer time-conditioned operations if the modulator waveform is shaped for multi-level or multi-bin control rather than a simple half-wave flip.","Extending parity labeling from qubits to time-bin qudits would require a larger modular spacing and multi-level modulators synchronized to that grid."],"forward_implications":["A deterministic routing gate between polarization and time-bin encodings on one photon becomes available for multi-DOF circuits such as polarization Toffoli constructions.","Either three-CNOT ordering can be realized by rearranging PBS-delay and EOM stages.","A concrete timing budget Δt ≳ τ_switch + τ_open + τ_det + σ_jitter sets the minimum viable bin spacing and overall gate duration.","With representative hardware numbers the paper quotes an ideal average gate rate near 400 MHz before loss and detector dead time."],"fun_headline_variants":["Parity time-bin encoding enables polarization-time qubit SWAP","Even-odd time bins turn a fixed delay into a logical bit flip","Parity labels let ordinary delays and modulators SWAP photon DOFs","Even/odd multiples of Δt unlock deterministic polarization-time SWAP","Parity-based time bins make three CNOTs a single-photon SWAP"],"cache_read_input_tokens":128,"weakest_assumption_plain":"That real modulators and delay stages can keep voltage accuracy and timing windows tight enough for polarization flips and bin assignments to stay correct through the full three-stage cascade.","fun_headline_variants_meta":{"raw":{"variants":["Parity time-bin encoding enables polarization-time qubit SWAP","Even-odd time bins turn a fixed delay into a logical bit flip","Parity labels let ordinary delays and modulators SWAP photon DOFs","Even/odd multiples of Δt unlock deterministic polarization-time SWAP","Parity-based time bins make three CNOTs a single-photon SWAP"]},"model":"grok-4.5","effort":"low","cost_usd":0.004586,"raw_usage":{"total_tokens":1352,"prompt_tokens":827,"num_sources_used":0,"completion_tokens":92,"cost_in_usd_ticks":45864000,"prompt_tokens_details":{"text_tokens":827,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":433,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":827,"tokens_out":92,"duration_ms":8354,"temperature":1.0,"reasoning_tokens":433,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T10:53:55.773598+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Build the PBS-delay / EOM / PBS-delay cascade at a chosen Δt, inject known polarization–time-bin product states, and test whether the output matches the swapped state within the paper’s predicted flip-error and jitter budgets; systematic mismatch beyond those models would falsify a working physical SWAP.","supporting_citations":[],"review_version":1}