{"id":"0e0427dc-4ba4-439d-ab0d-fb4d79642d39","arxiv_id":"2504.19490","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"sGA, a symmetry-constrained genetic algorithm that optimizes only the even-parity part of the wavefront, corrects two-photon correlations after a diffuser four times faster and up to 38% better than standard GA in one diffuser realization.","lead":"A new version of a genetic algorithm, called sGA, corrects distortions to pairs of entangled photons by optimizing only the symmetric half of the wavefront. It reaches the same correction roughly four times faster and yields a larger signal after a fixed number of steps.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"EMCCD contrast values in Fig. 3 contradict the claimed sGA superiority and must be reconciled.","rationale":"The paper's central claim is that enforcing even-parity symmetry makes the genetic algorithm both faster (4x) and better (38% higher enhancement). This claim rests on two experimental observables: the SPAD-based enhancement curves and the EMCCD-measured 2D correlation contrast. The enhancement curves are the fitness signal used by the optimizer, so they are not an independent validation of the final corrected state; the EMCCD correlation images are the independent check. The manuscript's reported contrast values (7.6 for GA, 5.5 for sGA) directly contradict the sentence claiming that Fig. 3(f) has higher contrast and, if taken literally, show GA outperforming sGA on the independent metric. That makes this the most load-bearing issue: a single transposition can be fixed, but if the values are accurate, the central claim is unsupported. The reader's identified weakest assumption (center alignment) is real but is acknowledged and probed in Appendix A, and the 20-pixel shift test shows only a gradual degradation; it does not contradict the aligned-case result. The single-disorder-realization limitation is also real but is a generality caveat, not an internal inconsistency. Thus the contrast contradiction is the sharpest concern. If the authors provide raw data and correct the inconsistency, the conditional acceptance can proceed; if the contradiction persists, the experimental evidence for sGA superiority collapses to the optimization-feedback curve alone, which is weaker. I therefore keep the reader's conditional verdict, with the condition extended to include a mandatory reconciliation of the contrast values.","tokens_in":9759,"tokens_out":9525,"duration_ms":103721,"concrete_test":"Recompute C = [mu(Peak) - mu(BG)] / sigma(BG) from the raw EMCCD frames underlying Fig. 3(e) and 3(f), using the exact peak region and background definition in Eq. (3), and report which panel corresponds to GA and which to sGA. If the recomputed contrast for the sGA panel is 5.5 while the GA panel is 7.6, the manuscript must either correct the figure/order or withdraw the statement that sGA yields higher-contrast correlations; if the recomputed sGA contrast is 7.6, correct the numbers in the text. Also verify that both contrasts were computed from the same number of frames and the same background mask, since the text says approximately 800,000 frames for panel (d) but only about 100,000 for panels (e) and (f).","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section IV reports, immediately after Eq. (3), that the EMCCD contrast for Fig. 3(e) and 3(f) is 7.6 and 5.5, respectively, and then states that Fig. 3(f) exhibits higher contrast. However, the figure order defined two paragraphs earlier is (e) = after standard GA and (f) = after sGA. With the stated numbers, the GA image has contrast 7.6 and the sGA image 5.5, i.e., the opposite of the claimed superiority. This is not a purely cosmetic error: the 2D correlation contrast is the independent, post-hoc measurement of correction quality, whereas the enhancement curves in Fig. 3(g) are built from the same SPAD coincidence signal that drives the optimization, so they are not fully independent evidence. If the numbers are simply transposed, the text and figure must be corrected; if they are not, the main claim that sGA outperforms GA is contradicted by the authors' own EMCCD data. Appendix B's analogous comparison (5.2 vs 5.4) shows only a marginal sGA advantage, making the large reversal in Fig. 3 particularly suspicious. The paper needs raw correlation images or recomputed contrast values to resolve this.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper demonstrates adaptive wavefront correction for spatially entangled photons using a genetic algorithm that enforces even-parity (180-degree rotational) symmetry on the SLM correction phase, based on the theoretical result that only the even-parity component of far-field diffuser phase affects two-photon correlations. The authors compare this symmetrized GA (sGA) with a standard GA using SPAD coincidence counts as feedback, reporting that sGA achieves the same enhancement in 25 generations that GA reaches in 100 generations, a claimed 4x speedup, and a 38% higher final enhancement at 100 generations. They also study the dependence of enhancement on detector integration time and include appendices on the sensitivity of sGA to center misalignment and detector displacement. The manuscript contains a direct internal inconsistency in the reported EMCCD contrast values, which the authors use as an independent verification of correction quality.","tokens_in":9984,"tokens_out":5034,"duration_ms":49758,"significance":"If the central claims survive scrutiny, the work is a useful practical advance: it uses direct quantum feedback rather than a classical probe beam, exploits a symmetry-based reduction of the optimization space, and includes robustness checks (center alignment, detector placement, integration time) that are often absent in wavefront-shaping papers. The theoretical motivation is clearly derived from Eq. (2) and the advanced-wave picture. However, the paper's own independent EMCCD contrast measurement currently contradicts the claimed sGA superiority, and the stated speedup factors are presented without the supporting numerical values and error bars. These issues must be resolved before the claims can be accepted.","major_comments":[{"comment":"The manuscript states that the contrast values for Fig. 3(e) and Fig. 3(f) are 7.6 and 5.5, respectively, and then concludes that Fig. 3(f) exhibits higher contrast. Since Fig. 3(e) is identified two paragraphs earlier as the result after standard GA and Fig. 3(f) as the result after sGA, the reported numbers imply that the GA image has higher contrast, directly contradicting the claim that sGA achieves superior performance. Because the EMCCD correlation measurement is the only independent, post-hoc check of correction quality (the enhancement curves in Fig. 3(g) use the same SPAD coincidence signal that drives the optimization), this inconsistency is load-bearing. The authors must either correct the transposed numbers or figure labeling, or provide the raw correlation images and recomputed contrast values that support the claimed ordering.","section":"Section IV, after Eq. (3)"},{"comment":"The sGA advantage relies on the optimization center (cx,cy) coinciding with the beam center (X,Y) on the SLM, yet the paper does not report the uncertainty of the beam-center measurement described in Fig. 2. The final scan in panels (e) and (j) uses r=20 and a cropped region, but no step size or repeated-measurement statistics are given. Given that Appendix A shows a 20-pixel shift (about half the disorder scale length of ~32 pixels) reduces the final sGA enhancement by roughly 40%, the authors should quantify the precision of their center-finding procedure and demonstrate that the Fig. 3 comparison is not degraded by a systematic center offset. Without this, the claim that sGA is four times faster is conditional on an unquantified alignment accuracy.","section":"Appendix A and Section III (beam center determination)"},{"comment":"The quantitative claims that sGA outperforms GA by 38% at the end of 100 generations and that sGA reaches GA's final enhancement of 5.9 in 25 generations are stated in the text without the corresponding numerical values or standard deviations from the five independent runs. The shaded error regions in Fig. 3(g) provide some information, but the 38% and 4x figures should be reported explicitly as mean ± standard deviation over the five runs, along with the actual enhancement values at generation 100 for both algorithms, to allow readers to assess the statistical significance of the claimed advantage.","section":"Section IV, Fig. 3(g)"}],"minor_comments":[{"comment":"The sentence 'calculated using data recoded by EMCCD' contains a typo: 'recoded' should be 'recorded'.","section":"Section IV, paragraph on EMCCD processing"},{"comment":"The exponential term is rendered as 'e−2k2 1σ 2−' and appears to have a formatting error; it should read e^{-2 k_1^2 \\sigma_-^2}, with the subscript 1 on the first k and the subscript - on sigma.","section":"Section II, Eq. (2)"},{"comment":"Capitalization of 'Fig.' is inconsistent; both 'Fig. 3' and 'fig. 3(e),(f)' appear in the same paragraph. Please standardize.","section":"Throughout"},{"comment":"References 14 and 15 are identical (both are Peeters, Moerman, and Van Exter, 'Observation of two-photon speckle patterns,' PRL 104, 173601 (2010)). One should be replaced with the correct citation if a different work was intended.","section":"References"},{"comment":"The expression 'ei(φ (k1)+φ (−k1))' has an extra closing parenthesis; the intended argument is φ(k1)+φ(-k1).","section":"Section II, sentence after Eq. (2)"}],"recommendation":"major_revision","confidential_remarks":"The experiment appears well designed and the appendices show a thoughtful treatment of alignment sensitivities. The single most important issue is the internal contradiction between the EMCCD contrast numbers and the claimed ordering of GA versus sGA performance; this must be resolved with the original data before publication. If the authors can supply corrected contrast values or raw correlation images, the paper is likely suitable for publication in a quantum-optics journal. I recommend requesting the raw EMCCD data and a detailed center-finding uncertainty analysis during the revision process."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid experiment showing that a symmetry-constrained genetic algorithm (sGA) restores two-photon correlations faster and with better enhancement than a standard GA. The main speedup claim—4x fewer generations to reach the same enhancement, and 38% better enhancement at fixed generation count—is supported by the evolution curves in Fig. 3(g), which are the primary evidence. The paper is worth a serious referee.\n\nWhat's genuinely new: earlier work (including the authors' own) established that only the even-parity component of a far-field diffuser affects two-photon correlations. Here they turn that into an optimization trick—enforce 180-degree rotational symmetry on the SLM phase, halving the search space—and demonstrate it experimentally in a high-dimensional SPDC setup. The appendices are a real strength: testing how the result degrades when the symmetry center is misaligned (Appendix A) and when the detectors are off-center (Appendix B) is exactly the kind of sensitivity analysis that makes me trust the main effect. The integration-time dependence is also a clean, useful observation.\n\nThe soft spots are real but not fatal. The most glaring is in Section IV: the text says the EMCCD contrast for Fig. 3(e) and (f) is 7.6 and 5.5, then claims (f) is higher. Unless the labels or numbers are transposed, that directly contradicts the stated conclusion. This matters because the EMCCD 2D correlation is the independent post-hoc measure of correction quality, whereas the enhancement curves come from the same coincidence feedback that drives the optimization. I'd want the raw correlation images or correctly labeled numbers before signing off. Also, the comparison rests on a single disorder realization. They run five independent optimizations on that one realization, which helps statistics but not generality. No code or raw data is provided, only \"available upon request,\" which is common in this field but still limits independent checking. The theory section is fine; the even-parity argument is clearly cited and the Klyshko picture makes it intuitive.\n\nBottom line: this is a practical advance for wavefront shaping with quantum light, not a paradigm shift. The reader's conditional verdict is about right. The fix for the contrast numbers is straightforward, and I'd like to see the authors respond to it, but the central claim is credible and the sensitivity analysis is honest. Send it to peer review.","headline":"A useful experimental demonstration that enforcing even-parity symmetry in a genetic algorithm speeds up wavefront correction for entangled photons, though a contradictory contrast number in the text needs to be reconciled before the secondary claim is clean.","tokens_in":10529,"tokens_out":1280,"would_cite":true,"duration_ms":15490,"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":"This paper shows that restricting wavefront corrections to the even-parity component of the phase distortion makes adaptive correction of spatially entangled photons four times faster and 38% more effective in a fixed number of generations.","keywords":["spatial entanglement","wavefront correction","symmetrized genetic algorithm","two-photon speckle","even-parity symmetry","adaptive optics","spontaneous parametric downconversion","beam center alignment"],"falsifier":"Measure the two-photon coincidence enhancement while applying a phase pattern that is purely odd under 180° rotation, centered on the beam; the theory says it should leave the correlation completely unchanged, so any measurable change would refute the premise behind the symmetrized algorithm.","tokens_in":9558,"feed_emoji":"🔬","tokens_out":8534,"duration_ms":75953,"temperature":0.7,"pith_summary":"Propagation through a random diffuser scrambles the transverse spatial correlations of entangled photon pairs, ruining the resource that quantum imaging and quantum communication rely on. This paper introduces a symmetrized genetic algorithm (sGA) that corrects the wavefront using only the even-parity part of the phase distortion, based on the proof that odd-parity distortions leave two-photon correlations untouched. Because the search space is halved, sGA reaches in 25 generations the enhancement that a standard genetic algorithm achieves in 100, and in the same 100 generations it delivers a 38% higher enhancement. The paper also shows that longer detector integration times improve the feedback signal-to-noise ratio and therefore the final enhancement. The practical payoff is faster, more robust wavefront correction of entangled light without needing an auxiliary classical beam.","feed_headline":"Symmetry makes wavefront correction 4x faster for entangled photons","feed_subtitle":"Even-parity-only correction halves the search space and boosts two-photon enhancement by 38%.","key_machinery":"The load-bearing object is the even-parity phase φ(k) + φ(−k) that appears in the two-photon correlation integral, together with the enforced 180°-rotational symmetry of the SLM phase mask that makes it the only degree of freedom. The identity that carries the argument is that the two-photon correlation C(x1,x2) depends on the diffuser only through the symmetrized combination, so an odd-parity phase mask is invisible to the feedback signal; this is what allows halving the parameter space. The mechanical enforcer is the sGA, a genetic algorithm in which the lower half of the super-pixel phase pattern is set by rotating the independently optimized upper half, plus a Zernike-based beam-centering scan that locates the symmetry center on the SLM.","core_discovery":"The central discovery is that the two-photon correlation function in the far field of a crystal is sensitive only to the even-parity component of a phase diffuser placed in the momentum plane, because the odd-parity component cancels in the advanced-wave picture. Starting from the SPDC two-photon wavefunction and the advanced-wave model, the authors show that the coincidence rate is governed by a correlation function into which the diffuser phase enters only through φ(k) + φ(−k). They exploit this by running a genetic algorithm on the SLM with the constraint that the phase mask be symmetric under 180° rotation about the beam center, so only half the super-pixels are independent. Experimentally, sGA achieves a fourfold speedup: an enhancement of 5.9 that takes the standard GA 100 generations is reached in 25 generations, and at 100 generations sGA exceeds GA by 38% in enhancement, with higher contrast in the restored two-photon correlation. The paper also establishes that the achievable enhancement rises with the signal-to-noise ratio of the coincidence feedback, which is set by the detector integration time.","pith_inferences":["A testable extension: restricting any wavefront-shaping optimizer (particle swarm, simulated annealing, or neural feedback) to phase masks with 180° rotational symmetry should yield a similar speedup, since the even/odd immunity is a property of the two-photon correlation, not of the genetic algorithm itself.","A consequence the paper does not draw: sGA's advantage over GA should grow with disorder strength, because stronger disorder makes the even-parity contribution more dominant and simultaneously sharpens the center-alignment sensitivity reported in Appendix A.","The Zernike-based beam-centering procedure exploits the same immunity and could serve as a general in situ alignment diagnostic for biphoton wavefront-shaping setups when no diffuser is present.","Because the 38% figure is measured with a single diffuser realization, averaging over many disorder realizations would likely reduce the margin; the paper's fixed-realization comparison is fair but may overstate the practical gain for arbitrary disorder."],"forward_implications":["Wavefront correction for spatially entangled photons can be performed with direct feedback from the quantum signal itself, without an auxiliary classical beam or pump-beam alignment, at up to four times lower overhead.","For a fixed optimization time, correcting only the even-parity component yields up to 38% higher enhancement of two-photon correlations than correcting the full wavefront.","The speed advantage depends on accurate localization of the beam center on the SLM: a center offset of 20 pixels (about half the disorder scale length) reduces sGA's final enhancement by about 40%, while GA is unaffected by such shifts.","Because the odd-parity component does not contribute, detectors must be positioned at the beam center; placing them off-center removes sGA's advantage, since sGA cannot add the beam tilt that GA uses to improve feedback.","Longer detector integration times raise the signal-to-noise ratio of the coincidence feedback and increase the final enhancement for both algorithms, at the cost of longer total runtime."],"supporting_citations":[{"why":"Provides the biphoton wavefunction and the transverse spatial correlation formalism for SPDC that the derivation starts from.","marker":"1"},{"why":"Reports two-photon speckle patterns and supplies the correlation function that motivates the even-parity decomposition.","marker":"14"},{"why":"Demonstrates real-time shaping of entangled photons via pump-beam feedback, the prior method this work extends by using direct quantum feedback without an auxiliary beam.","marker":"21"},{"why":"Establishes partial immunity of two-photon correlations to wavefront distortion, the even/odd decomposition insight on which sGA is built.","marker":"23"},{"why":"Presents the advanced-wave picture used to explain why odd-parity phase cancels in the two-photon measurement.","marker":"25"},{"why":"Applies the advanced-wave picture to shaping through scattering media, supporting the symmetry argument.","marker":"26"},{"why":"Supplies the genetic-algorithm optimization methodology (population, mutation, elites) used by both GA and sGA.","marker":"34"}],"fun_headline_variants":["Symmetry halves search space, speeds entangled photon correction","Even-parity trick accelerates quantum wavefront fixing 4x","Genetic algorithm with symmetry fixes entangled photons faster","Halving parameters speeds up entangled photon wavefront correction","sGA halves search space, restores entangled photon correlations faster"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The speed and enhancement gains of the symmetrized algorithm depend on the correction pattern being centered on the exactly right spot of the light beam; if that center is off by about half the size of the random variations, the improvement drops by roughly 40%.","fun_headline_variants_meta":{"raw":{"variants":["Symmetry halves search space, speeds entangled photon correction","Even-parity trick accelerates quantum wavefront fixing 4x","Genetic algorithm with symmetry fixes entangled photons faster","Halving parameters speeds up entangled photon wavefront correction","sGA halves search space, restores entangled photon correlations faster"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000588,"raw_usage":{"total_tokens":2744,"prompt_tokens":909,"completion_tokens":1835,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":525,"completion_tokens_details":{"reasoning_tokens":1757}},"tokens_in":525,"tokens_out":1835,"duration_ms":13887,"temperature":1.0,"reasoning_tokens":1757,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:51:10.393887+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the two-photon coincidence enhancement while applying a phase pattern that is purely odd under 180° rotation, centered on the beam; the theory says it should leave the correlation completely unchanged, so any measurable change would refute the premise behind the symmetrized algorithm.","supporting_citations":[{"cited_title":"Lib \\ and\\ author Y","cited_arxiv_id":null,"evidence_quote":"Demonstrates real-time shaping of entangled photons via pump-beam feedback, the prior method this work extends by using direct quantum feedback without an auxiliary beam."}],"review_version":1}