{"id":"921a14b9-240d-4b92-a3b3-671ab4bffe25","arxiv_id":"2602.09495","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Applying the NulLA algorithm proves no-go theorems for heralded linear-optical state generation, showing e.g. that Bell states need ≥4 input photons and a heralded CNOT gate needs ≥2 ancilla photons.","lead":"This paper uses algebraic geometry to rigorously prove that certain photonic quantum states and gates cannot be created with fewer photons than known schemes use. It gives a way to turn 'we couldn't find a solution' into 'no solution exists' for linear-optical state generation.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 1's reduction to all-ones herald patterns is unproven and likely false for zero-photon herald modes; no-go generality is conditional.","rationale":"The reader identified Lemma 1's optimal-configuration reduction as the weakest load-bearing premise, and I agree. The central claim that NulLA certificates constitute rigorous no-go theorems for all configurations with given photon numbers depends on Lemma 1. The proof of that lemma has a concrete logical gap: the fan-out argument mishandles incorrect herald outcomes with total photon number greater than m, and it does not address zero-photon herald modes, which are used in the paper's own vacuum-heralded examples. Since the lemma is stated in full generality but proved only for positive-count herald patterns, the sweeping lower bounds for Bell states and the CNOT gate are not fully established. This does not require rejecting the paper: the NulLA certificates themselves are valid for the exact configurations tested, and the gap may be repairable by adding a careful argument about total photon-number conservation and by either excluding zero-count patterns from the lemma's scope or proving a separate reduction for them. The reader's CONDITIONAL verdict already captures this, so my stress-test does not change it.","tokens_in":15378,"tokens_out":24621,"duration_ms":240737,"concrete_test":"Run the Bell-state no-go test from Sec. IV.A with N=5, n=3, m=1, but replace the herald pattern (1) with (0,1) on two herald modes (same total m=1), keeping target and input identical. The paper reports a degree-9 certificate for (1). If the (0,1) system has no degree-9 certificate and a Groebner-basis check finds a solution, Lemma 1 is false; if it also returns a certificate, the lemma may be repairable and the conditional can be lifted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section III.A's Lemma 1 is what turns individual NulLA runs into no-go theorems for all separable n-photon inputs and all m-photon heralding patterns. The proof sketch is incomplete. The fan-out step asserts that 'any incorrect distribution must have fewer photons than (m1,...,mM) in at least one of the modes,' which is false for outcomes whose total photon number exceeds m (e.g., desired (1,2), incorrect (2,3)); rejecting those requires the additional—unstated—assumption that all fan-out output modes are measured by photon-number-resolving detectors and that the all-ones pattern has exactly m modes. More seriously, the reduction only treats positive m_j. It cannot simulate a herald pattern containing zero-photon modes, such as (0,1) or the vacuum-heralded schemes used in Sec. IV.C, because linear optics cannot turn a vacuum requirement into the presence of a photon in an all-ones pattern. Lemma 1 is stated for 'all heralding patterns containing m photons,' so Theorem 2 and the Bell/CNOT lower bounds inherit this gap. Absent a proof covering zero-count patterns, the certificates are rigorous only for the precise configurations run, not for the claimed all-configuration lower bounds.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper proposes to use the Nullstellensatz Linear Algebra (NulLA) algorithm to prove infeasibility of heralded linear-optical state generation. Following the polynomial formulation of Ref. [17], the state generation task is encoded as a system of polynomial equations in the entries of an unknown linear transformation A and a scaling parameter gamma; a NulLA certificate 1=sum beta_i f_i definitively proves that no such transformation exists. The authors introduce a simplification lemma (Lemma 1) claiming that, for fixed total photon numbers, only the single-photon-per-mode input and the all-ones heralding pattern need be tested. They report infeasibility certificates for Bell state generation with three input photons, for random two-photon target states, for vacuum-heralded NOON states from non-single-photon inputs, and for a one-ancilla-photon CNOT gate, and hence claim lower bounds on resource requirements such as at least four photons for a Bell state and at least two ancilla photons for a heralded CNOT.","tokens_in":15678,"tokens_out":16093,"duration_ms":168301,"significance":"If the central reduction and the certificates are valid, the paper offers a genuinely useful tool: NulLA provides rigorous, machine-checkable no-go statements where previous work often relied on numerical search, and the reported low certificate degrees relative to worst-case bounds are encouraging. The authors also correctly avoid the equivalence-class shortcut that can cause false infeasibility (Appendix A), and they make code available. The specific results--Bell state requiring four photons, CNOT requiring two ancilla photons, and NOON states requiring single-photon-per-mode inputs for vacuum-heralded generation--would be valuable lower bounds for photonic resource analysis. However, the generality of these results currently depends on Lemma 1, whose proof is incomplete; as written, the certificates are rigorous only for the precise test configurations, not for the advertised all-configuration theorems.","major_comments":[{"comment":"The herald-pattern reduction is not proven. The proof asserts that any incorrect distribution has fewer photons than (m_1,...,m_M) in at least one mode; this is false for outcomes whose total photon number exceeds m, e.g. desired (1,2) and incorrect (2,3). Such outcomes could still trigger the all-ones pattern after fan-out unless extra total-photon-number or PNR conditions are imposed. More importantly, the fan-out construction cannot map a heralding pattern with a zero-photon mode (e.g. (0,1), or the m=0 vacuum-heralded case in Sec. IV.C) to the all-ones pattern, because passive linear optics cannot create a photon from vacuum. The input-reduction step itself produces vacuum-herald events, and these are not handled. Consequently Theorem 2 and the lower bounds for all separable n-photon inputs and all m-photon heralding patterns are not established; the certificates are rigorous only fo","section":"Section III.A, Lemma 1 and Theorem 2"},{"comment":"The treatment of non-unitary transformations is under-specified. The text states that a general solution A can be rescaled so that ||A||<=1 and then embedded in a larger unitary, but it does not explain what happens to the additional modes in the heralded setting. If the dilation leaves photons in the added modes, they must be measured or postselected, typically to vacuum, which adds herald modes that are absent from the polynomial system (2.8). This matters because the added vacuum conditions interact with the unproven zero-mode part of Lemma 1. Please either give the standard embedding argument for heralded generation, with the fate of the extra modes, or cite the precise result in Refs. [17,25].","section":"Section II.A (after Eq. 2.4)"},{"comment":"The tested system is described as the three-photon, five-mode transformation derived from the original scheme when one input photon is removed. If this is only a specific ansatz obtained by deleting a photon from the known two-ancilla scheme, the certificate rules out that ansatz, not all one-heralding-photon CNOT gates. The text should clarify that the polynomial system solved is the fully general 3-photon, 5-mode herald-one transformation with the four logical-basis input-output pairs; otherwise the conclusion that two ancilla photons are necessary is not supported.","section":"Section IV.D.1 (CNOT)"},{"comment":"The no-go results are computer-assisted, but the actual Nullstellensatz certificates (the polynomials beta_i) are not included; only the certificate degrees are reported. For a rigorous, checkable proof, the certificates or a permanent, versioned data/code release that verifies the identities 1=sum beta_i f_i should be provided. The informal GitHub URL is not sufficient for peer review. This is a reproducibility issue, not an issue with the NulLA method itself.","section":"Sections IV.A-IV.D and Ref. [44]"}],"minor_comments":[{"comment":"The phrase 'all separable n-photon input states' should be narrowed to 'all product Fock states with n photons' or explicitly defined, to avoid confusion with superpositions of Fock states.","section":"Section III.A"},{"comment":"Please define V and s in the displayed Kollar bound; currently they are introduced only in the following sentence.","section":"Section III.B, Eq. (3.3)"},{"comment":"In the row with no certificate, replace the dash in the 'Computed degree' column with 'none up to d=9'.","section":"Table I"},{"comment":"The URL contains spaces and is not stable; provide a permanent repository identifier such as a DOI or a specific commit hash.","section":"Ref. [44]"},{"comment":"The captions should state what the detector patterns represent (photon-number-resolving outcomes) and explain the dashed modes; the green/red colour coding is hard to decode from the current captions alone.","section":"Figures 1 and 2"}],"recommendation":"major_revision","confidential_remarks":"The main technical gap is Lemma 1; if the authors can prove it under clearly stated detector assumptions, the paper could be acceptable after revision. The unitarization point should also be addressed. I do not see grounds for rejection, as the method and specific certificates are valuable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid methods paper that gives rigorous no-go certificates for heralded linear-optical state generation, and the main theorem is more likely right than wrong, but the proof of Lemma 1 as written has a hole that should be fixed before publication.\n\nThe genuinely new thing is using NulLA instead of Gröbner bases or numerical search to decide infeasibility of the polynomial system that encodes a heralded state-generation task. That yields definitive lower bounds: Bell states need at least four input photons, a heralded CNOT needs at least two ancilla photons, and NOON states with vacuum heralding need the input photons spread one per mode. Those are useful, concrete results. The Bell bound was known, but the CNOT and NOON bounds are new. The random-state sweeps show that low-degree certificates are typical, which is evidence that the method is practical.\n\nWhat is good: the reduction to the polynomial system is standard and the NulLA certificates are rigorous proofs. The paper is honest about the difference between finding a certificate and exhausting the degree search. The scaling bounds are helpful.\n\nThe soft spot is Lemma 1. The claim that the all-ones herald pattern suffices is proved by a fan-out argument that says any incorrect distribution must have fewer photons in at least one mode. That is false for outcomes with more total photons than the correct pattern — e.g., (1,2) vs (2,3). The argument can be fixed by noting that the all-ones pattern has exactly m modes and requires exactly m photons, so any outcome with the wrong total photon number is automatically rejected. But that is not stated, and a reader cannot fill the gap from the text. The same lemma also glosses over zero-photon herald modes; the intended claim is that the optimal pattern has m entries, so vacuum modes are simply dropped, but the proof sketch doesn't say that explicitly. I don't think these are fatal — the reduction is likely correct — but the proof needs to be written out properly if the no-go theorems are to be as general as advertised.\n\nA minor issue: the GitHub link looks like a placeholder and the actual certificates are not in the arXiv. A referee should ask for the code and at least one certificate to verify the computational claims.\n\nRecommendation: send it out. The method is useful and the gaps are repairable. I would ask for a rigorous rewrite of Lemma 1 and an accessible code link before acceptance.","headline":"A useful NulLA-based method for rigorous photonic resource lower bounds; Lemma 1 is under-proved but the reduction is likely salvageable.","tokens_in":16156,"tokens_out":12731,"would_cite":true,"duration_ms":86561,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Applying the Nullstellensatz Linear Algebra algorithm to the polynomial equations of heralded photon circuits yields rigorous impossibility proofs and certified lower bounds on photon resources.","keywords":["heralded state generation","linear optics","Nullstellensatz","infeasibility certificates","no-go theorems","Bell states","CNOT gate","NOON states"],"falsifier":"Find an explicit linear-optical network that heralds a Bell state from a three-photon separable input (contradicting the degree-9 certificate), or a CNOT gate from a single ancilla photon (contradicting the degree-6 certificate). Alternatively, construct a separable n-photon input distribution with more than one photon in some mode that can be heralded to produce a target state even though the single-photon-per-mode configuration with the same n is infeasible; that would disprove Lemma 1.","tokens_in":15298,"feed_emoji":"⚛️","tokens_out":4699,"duration_ms":47708,"temperature":0.7,"pith_summary":"The central claim is that the feasibility of a heralded linear-optical state-generation task can be decided by asking whether a system of polynomial equations has a solution, and that the NulLA algorithm can certify 'no solution' with mathematical certainty. Such a certificate is stronger than a failed numerical search: it proves that no linear-optical network, however clever, can turn the given input into the target state under the given heralding pattern. Using this, the paper establishes that a Bell state cannot be produced from three separable input photons (so four are necessary), that a heralded CNOT gate requires at least two ancilla photons, and that vacuum-heralded NOON states require single-photon-per-mode inputs. The paper also proves a simplification lemma that reduces the infinite family of possible input and heralding configurations to one canonical configuration per total photon number, making the tests feasible. A sympathetic reader would care because these are the first rigorous resource lower bounds of this kind for ubiquitous photonic building blocks.","feed_headline":"Algebra turns 'no scheme found' into proof a Bell state needs four photons","feed_subtitle":"A Nullstellensatz check on polynomial equations certifies lower bounds for Bell, CNOT, and NOON states.","key_machinery":"The central object is the Nullstellensatz certificate: polynomials β_i such that 1 = Σ β_i f_i over the polynomial ring in the entries of the transformation matrix A. By Hilbert's Nullstellensatz, existence of such a certificate is equivalent to the system f_1=...=f_s=0 having no solution over the complex numbers. NulLA turns the search for a certificate of a fixed degree d into a linear algebra problem, and the paper's Lemma 1 reduces the search space: to rule out all separable n-photon inputs and all m-photon heralding patterns, it suffices to test the input |1,1,...,1> and the herald (1,1,...,1). The polynomial encoding of state generation follows an existing construction: the input Fock","core_discovery":"The discovery is that the NulLA algorithm, a standard tool from algebraic geometry for proving polynomial systems infeasible, applies directly to heralded linear optics. The state-generation condition — that the heralded output equals the target up to scaling — is rewritten as γG = Q, where G and Q are homogeneous polynomials in creation operators with coefficients depending on the unknown linear transformation A. Equating monomial coefficients gives a system of polynomial equations in the entries of A; if these equations have no solution, the task is impossible. NulLA searches for a Nullstellensatz certificate of increasing degree, and finding one is a rigorous proof of infeasibility. The p","pith_inferences":["If Lemma 1's fan-out proof is completed (the written argument leaves the case of outcomes with more than the herald photon count implicit), the lower bounds would rigorously cover all separable inputs with the same total photon number; as written they cover the canonical configuration plus whatever the reduction genuinely establishes.","The same certificate machinery could be applied to other resource questions, such as the minimum photon number for fusion gates or for post-selected state generation, by writing the appropriate polynomial systems; the paper does not address these.","The observation that certificate degrees stay low across random targets hints at a general phenomenon — the effective Nullstellensatz degree for these photonic systems may be far below Kollár's bound — which, if true, would make NulLA a practical decision procedure for a wide range of photonic tasks.","A repaired or strengthened Lemma 1 would also open the door to classifying tasks by photon number alone, effectively creating a lookup table of feasible vs infeasible state-generation tasks; the current paper stops at demonstrating examples."],"forward_implications":["Bell-state generation from three separable photons is impossible; the known four-photon schemes are resource-optimal.","A heralded CNOT gate needs at least two ancilla (heralding) photons; single-ancilla versions are ruled out.","Vacuum-heralded NOON states cannot work if any input mode contains more than one photon; the standard single-photon-per-mode inputs are therefore necessary.","For families of random two-photon target states, infeasibility certificates appear at fixed low degrees, suggesting that low-degree certificates are typical rather than exceptional.","The method extends from fixed Fock inputs to probabilistic sources by treating each photon-number sector as a separate task that a single transformation must satisfy."],"fun_headline_variants":["Algebra certifies photon-count floor for Bell, CNOT, NOON","No-go theorem: linear optics needs these many photons","Polynomial infeasibility proves quantum state prep limits","NulLA finds rigorous photon minimums for heralded gates"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The fan-out argument in Lemma 1, which asserts that any incorrect photon distribution must have fewer photons than the target herald pattern in at least one mode, assumes that detection outcomes with too many photons are excluded; if that step fails, the no-go theorems apply only to the tested configurations, not to all configurations with the same photon numbers.","fun_headline_variants_meta":{"raw":{"variants":["Algebra certifies photon-count floor for Bell, CNOT, NOON","No-go theorem: linear optics needs these many photons","Polynomial infeasibility proves quantum state prep limits","NulLA finds rigorous photon minimums for heralded gates"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000902,"raw_usage":{"total_tokens":3709,"prompt_tokens":721,"completion_tokens":2988,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":465,"completion_tokens_details":{"reasoning_tokens":2931}},"tokens_in":465,"tokens_out":2988,"duration_ms":24831,"temperature":1.0,"reasoning_tokens":2931,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T02:56:49.391675+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find an explicit linear-optical network that heralds a Bell state from a three-photon separable input (contradicting the degree-9 certificate), or a CNOT gate from a single ancilla photon (contradicting the degree-6 certificate). Alternatively, construct a separable n-photon input distribution with more than one photon in some mode that can be heralded to produce a target state even though the single-photon-per-mode configuration with the same n is infeasible; that would disprove Lemma 1.","supporting_citations":[],"review_version":1}