{"id":"d377f8ce-2cb6-43a8-9d4d-3e8fab422a56","arxiv_id":"2603.20139","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Two-mode squeezed light with two-port balanced homodyne yields an analytical classical Fisher matrix that Heisenberg-scales all four U(2) network parameters and saturates multiparameter CRBs via MLE at modest sample size.","lead":"A two-mode squeezed probe plus balanced homodyne at both outputs can estimate all four real parameters of a two-channel optical network at Heisenberg scaling. The scheme is fully Gaussian and aims at practical calibration of photonic devices and distributed sensors.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified beyond the incomplete extract already flagged by the reader.","rationale":"The reader’s strongest claim accurately restates the paper’s headline result, and the weakest assumption correctly names the load-bearing premise (full-rank Heisenberg classical FI for the entire four-parameter vector under two-port HD). The incomplete extract prevents independent verification of the analytic FI and the MLE saturation plots, which is precisely why the reader already assigned CONDITIONAL / MODERATE. No additional technical soft spot—such as an unstated purity assumption that would break under loss, a coordinate singularity that permanently nulls one parameter, or a mismatch between the claimed Gaussian likelihood and the actual homodyne statistics—appears in the available text. Therefore the stress-test does not alter the verdict; it simply reconfirms that the decisive check is the missing analytic FI matrix itself.","tokens_in":5331,"tokens_out":565,"duration_ms":6132,"concrete_test":"Once the full arXiv PDF is available, extract the analytic classical FI matrix (claimed after Fig. 1) and evaluate its four eigenvalues (or the diagonal of the inverse) at a generic point ϕ and at the singular loci of U(2) (e.g., ϕ3 = 0 or π/2) for several fixed (θ1, θ2). Confirm that every eigenvalue scales as ~N^{2} (Heisenberg) and that the matrix remains invertible; if any eigenvalue drops to O(N) or zero for all LO choices, the simultaneous four-parameter claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim (TMSS + two-port balanced HD yields a full-rank classical FI matrix with simultaneous Heisenberg scaling for all four U(2) parameters, and MLE saturates the multiparameter CRB at modest N and sample size) is consistent with the abstract, Fig. 1 setup, and the authors’ prior multiparameter squeezed-light results. The reader’s weakest assumption correctly isolates the key premise: that free choice of LO phases θ1, θ2 keeps the classical FI full-rank and ~N^{2} across the generic U(2) manifold, including the overall phase ϕ0 once the LO supplies a reference. Nothing in the supplied text contradicts that premise or reveals an internal inconsistency (e.g., a singular FI block, a hidden single-parameter restriction, or a failure of the Gaussian likelihood). The only genuine limitation is the incomplete manuscript extract itself—technical FI derivation, explicit matrix elements, and MLE numerics are absent—so the claim cannot be fully audited here. That is already the basis of the reader’s CONDITIONAL verdict; no stronger load-bearing flaw is visible.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript proposes a fully Gaussian multiparameter metrology scheme for the simultaneous estimation of all four real parameters of an arbitrary two-channel unitary U(ϕ) (a general U(2) linear-optical network). A two-mode squeezed vacuum probe is sent through the unknown network and both output ports are measured by balanced homodyne detection with freely chosen local-oscillator phases θ1, θ2. The authors claim an analytic derivation of the complete classical Fisher-information matrix for the four-parameter vector, simultaneous Heisenberg (1/N) scaling for every parameter, and numerical evidence that a maximum-likelihood estimator saturates the multiparameter Cramér–Rao bounds already for modest sample sizes (~100 repetitions) and low mean photon number. The work is presented as a practical route to full characterization of two-channel photonic devices.","tokens_in":5590,"tokens_out":840,"duration_ms":8052,"significance":"If the analytic classical Fisher matrix is indeed full-rank and scales as N^{2} for all four U(2) parameters simultaneously, and if the MLE saturation holds under realistic conditions, the result would supply a concrete, experimentally accessible protocol for Heisenberg-scaling multiparameter Gaussian metrology of arbitrary two-channel networks. That would be directly useful for calibration and sensing in integrated photonics and for distributed continuous-variable sensor networks. The scheme builds cleanly on the authors’ earlier multiparameter squeezed-light results while adding the two-port homodyne analysis and the finite-sample MLE study; those additions are the genuine technical contribution.","major_comments":[{"comment":"The supplied manuscript extract contains only the abstract, introduction, Fig. 1 caption, a brief closing paragraph on MLE saturation, acknowledgments and references. The technical body that must contain the analytic classical Fisher-information matrix, its explicit elements, the proof of simultaneous Heisenberg scaling for all four parameters (including the overall phase ϕ0 once the LO supplies a reference), and the MLE numerics is absent. Without those derivations the central claims cannot be audited; the paper cannot be accepted until the complete technical sections are provided and verified.","section":null},{"comment":"Even after the missing sections are restored, the claim that free choice of LO phases θ1, θ2 keeps the classical FI full-rank and ~N^{2} across the entire generic U(2) manifold (including ϕ0) remains the load-bearing premise. The final manuscript must exhibit the FI matrix (or its eigenvalues / determinant) as a function of ϕ and of the LO phases, and must demonstrate that there exist accessible LO settings for which no singular blocks appear for any of the four parameters.","section":null}],"minor_comments":[{"comment":"Title and abstract wording differ slightly (“arbitrary two-channel network” vs. “two-channel optical network”); unify for consistency.","section":null},{"comment":"Fig. 1 caption is clear, but the main text should explicitly state how the overall phase ϕ0 becomes identifiable once the local oscillators provide a phase reference.","section":null},{"comment":"References [23] and [29] are the authors’ own closely related works; a short paragraph clarifying the precise technical advance relative to those papers would help the reader.","section":null}],"recommendation":"major_revision","confidential_remarks":"The incomplete extract makes a definitive technical audit impossible; the CONDITIONAL verdict of the reader is therefore correct. Once the full technical body is supplied the paper may well be sound, but until then major revision (or resubmission of the complete manuscript) is the only responsible recommendation. Scope and novelty appear appropriate for a quant-ph letter once the missing material is restored."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The one thing worth knowing is that this is a concrete, all-Gaussian recipe (two-mode squeezed vacuum plus balanced homodyne at both ports) that claims a full-rank classical Fisher matrix with simultaneous 1/N scaling for every real parameter of a generic two-channel unitary, including the overall phase once the local oscillators supply a reference. That is useful for device calibration and continuous-variable multiparameter sensing, not a foundational breakthrough.\n\nWhat is actually new relative to the authors’ own 2025 PRA papers is the two-port homodyne analysis itself: they derive the complete classical FI matrix analytically for the four-parameter U(2) and show, via maximum-likelihood estimation, that the multiparameter CRBs are already saturated at roughly 100 shots and low mean photon number. The setup is experimentally realistic, the Gaussian likelihood is exact, and the free LO phases \theta1, \theta2 are treated as tunable resources that keep the matrix full-rank across the manifold. The citation pattern is honest—self-citation is present but points to the natural prior steps in the same program rather than circular re-labeling.\n\nThe soft spot is simply that the extract we have stops short of the technical body. The abstract, figure caption, and closing MLE paragraph state the claims cleanly, but the explicit matrix elements and the numerical saturation plots are missing, so we cannot audit the algebra or the finite-sample behavior ourselves. Nothing in the available text contradicts the central premise or shows a singular FI block; the stress-test correctly finds no stronger load-bearing flaw. Circularity is mild.\n\nThis paper is for people already working on continuous-variable multiparameter metrology or photonic-network characterization who want a practical, fully Gaussian route rather than photon-counting or non-Gaussian probes. It deserves a serious referee. I would send it to peer review; the result is incremental but cleanly framed and potentially useful once the missing derivation is checked.","headline":"Solid, incremental Gaussian protocol that claims simultaneous Heisenberg scaling for all four U(2) parameters via TMSS + two-port homodyne, with analytical classical FI and early MLE saturation; incomplete extract is the only real audit limit.","tokens_in":6221,"tokens_out":498,"would_cite":false,"duration_ms":4106,"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":"A two-mode squeezed probe and two-port homodyne detection can estimate all four parameters of a two-channel optical network at Heisenberg scaling at once.","keywords":["multiparameter quantum metrology","Heisenberg scaling","two-mode squeezed state","homodyne detection","Gaussian quantum optics","two-channel unitary","Cramér–Rao bound","integrated photonics"],"falsifier":"Compute or measure the classical Fisher matrix of the two-port homodyne statistics for a known two-channel unitary and check whether all four diagonal entries continue to scale as 1/N while the matrix remains invertible; any systematic drop below Heisenberg scaling or loss of rank would falsify the claim.","tokens_in":6180,"feed_emoji":"🔬","tokens_out":653,"duration_ms":4778,"temperature":0.7,"pith_summary":"This paper shows that a fully Gaussian optical setup can characterize an arbitrary two-channel linear network completely, and at Heisenberg precision, in a single multiparameter experiment. A two-mode squeezed vacuum is sent through the unknown unitary, after which balanced homodyne detection is performed at both output ports with freely chosen local-oscillator phases. The authors derive the classical Fisher-information matrix analytically and show that every one of the four real parameters of the network scales as 1/N with mean photon number N. Maximum-likelihood estimation is then shown to reach the multiparameter Cramér–Rao bounds already with roughly a hundred experimental runs and only a few photons. The result matters because full device calibration and multiparameter sensing in integrated photonics have until now faced trade-offs that prevent simultaneous Heisenberg scaling for all parameters; the scheme removes that barrier with experimentally standard continuous-variable tools.","feed_headline":"Four network parameters estimated at once at Heisenberg scale","feed_subtitle":"Two-mode squeezing plus dual homodyne saturates multiparameter bounds with few photons","key_machinery":"The analytically derived 4×4 classical Fisher-information matrix of the two-port homodyne outcomes, which is full rank and Heisenberg-scaling for every component of the parameter vector ϕ once the two local-oscillator phases are chosen freely.","core_discovery":"A pure two-mode squeezed probe measured by balanced homodyne detection at both output ports yields a full-rank classical Fisher-information matrix that simultaneously attains Heisenberg scaling (1/N) for all four real parameters of an arbitrary two-channel unitary U(ϕ). Maximum-likelihood estimation saturates the corresponding multiparameter Cramér–Rao bounds already for modest sample size and low mean photon number.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Dual homodyne on two-mode squeezed light hits Heisenberg for all four network params","Simultaneous Heisenberg-scale estimation of four two-channel optical network parameters","TMSV plus two-port homodyne saturates multiparameter CRB at 1/N for full U(ϕ)","Gaussian two-mode probe fully characterizes two-channel network at Heisenberg scaling","Four real parameters of two-channel unitary estimated at once with low-photon Heisenberg"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The scheme assumes ideal balanced homodyne detection of a pure two-mode squeezed vacuum, with free choice of local-oscillator phases, is enough to keep the classical Fisher matrix full-rank and Heisenberg-scaling for every parameter of a generic two-channel unitary.","fun_headline_variants_meta":{"raw":{"variants":["Dual homodyne on two-mode squeezed light hits Heisenberg for all four network params","Simultaneous Heisenberg-scale estimation of four two-channel optical network parameters","TMSV plus two-port homodyne saturates multiparameter CRB at 1/N for full U(ϕ)","Gaussian two-mode probe fully characterizes two-channel network at Heisenberg scaling","Four real parameters of two-channel unitary estimated at once with low-photon Heisenberg"]},"model":"grok-4.5","effort":"low","cost_usd":0.007208,"raw_usage":{"total_tokens":1734,"prompt_tokens":703,"num_sources_used":0,"completion_tokens":115,"cost_in_usd_ticks":72080000,"prompt_tokens_details":{"text_tokens":703,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":916,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":703,"tokens_out":115,"duration_ms":9161,"temperature":1.0,"reasoning_tokens":916,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-13T21:40:42.490176+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Compute or measure the classical Fisher matrix of the two-port homodyne statistics for a known two-channel unitary and check whether all four diagonal entries continue to scale as 1/N while the matrix remains invertible; any systematic drop below Heisenberg scaling or loss of rank would falsify the claim.","supporting_citations":[],"review_version":1}