{"id":"d5e05293-fb05-4844-8478-3d009c45887b","arxiv_id":"2608.09509","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A gate-based circuit with one qubit per optical mode plus one ancilla reproduces two-photon boson-sampling statistics, demonstrated for four modes.","lead":"This paper shows a recipe for turning a two-photon light-interference experiment into a circuit of quantum gates, and demonstrates it on a four-qubit photonic setup. The result is a small but working bridge between boson sampling and general-purpose quantum computers.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Ancilla condition in Sec. IV.B only triggers for local two-photon states; single-photon components of two-photon inputs (e.g., |1010>) are left unchanged, so the claimed input-independent N+1-qubit circuit is not a faithful boson-sampling emulation.","rationale":"The reader's weakest_assumption targets the missing general m-photon framework, but the more immediate problem is that the two-photon N-mode construction itself is incorrect as written. The load-bearing condition for the central claim is that the repeating unit faithfully implements a beam splitter within the two-photon Hilbert space. That requires acting on single-photon components as well, which the ancilla condition explicitly excludes. This is a correctness error, not just an overclaim about scalability. It also explains why the paper resorts to input-specific circuits: the general circuit would not match Eq. (3). The experimental fidelities are real evidence for the optimized circuits, but they cannot rescue the claimed scalable framework. The verdict should move from CONDITIONAL to REJECT unless the theoretical construction is substantially corrected.","tokens_in":20364,"tokens_out":15552,"duration_ms":135290,"concrete_test":"Take the four-mode interferometer U in Eq. (3) and the construction of Secs. IV–V. Encode the input |1010> as the computational-basis state |1010> on (q0,q1,q2,q3) plus ancilla. Propagate this state through the input-independent circuit of Fig. 6 exactly as specified by the repeating-unit rule: for each BS (1,2), (3,4), (2,3), compute the ancilla a = AND of NOTs of all spectator qubits, apply the ancilla-controlled two-photon gate, then uncompute. Compare the resulting output distribution with the exact two-photon probabilities for |1010> (column 6 of P_th: 0.125 each for |0200>, |0020>, |1100>, |1010>, |0101>, |0011> and 0.25 for |1001>). If the circuit leaves |1010> unchanged or produces any distribution other than P_th, the general construction is falsified. An alternative check: verify unitarity of the effective 10x10 matrix obtained by propagating all ten basis states through Fig.","verdict_should_be":"REJECT","load_bearing_attack":"The central theoretical claim is the input-independent (N+1)-qubit circuit for two-photon N-mode boson sampling. The construction replaces each beam splitter on (i,j) with a repeating unit whose ancilla is activated only when every spectator mode is unoccupied (Sec. IV.B: a = AND_{k≠i,j} NOT q_k). But a beam splitter acts on any photon in modes i,j, regardless of where the other photon is. For a two-photon input with one photon in mode i and one in spectator mode k, the local state is |1_i,0_j>; the corresponding single-photon beam-splitter transformation must produce a superposition including |0_i,1_j>, while the spectator photon stays in k. The ancilla condition is false (q_k=1), so the repeating unit applies no gate and the state passes through unchanged. This omission affects the first two beam splitters in the four-mode example: input |1010> would remain |1010> instead of evolving into the seven-component distribution given by column 6 of Eq. (3). Consequently, the generalized circuit in Fig. 6 cannot reproduce the two-photon evolution 'over the complete ten-dimensional Hilbert space' as claimed. The experimental section evades this because it uses input-specific optimized circuits (Fig. 7), not the general repeating-unit architecture, so the high fidelities do not validate the scalable mapping.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a gate-based qubit framework for emulating boson sampling. It constructs balanced beam-splitter circuits for the single-photon and two-photon sectors, combines them through an ancilla-assisted encoding, and claims a scalable repeating-unit architecture that maps a two-photon N-mode interferometer to an (N+1)-qubit circuit. A four-mode two-photon circuit is then developed, and input-specific optimized versions are implemented experimentally using a heralded single photon encoded in path and polarization. The experimental output distributions are compared with permanent-based theoretical predictions, reporting squared classical fidelities above 0.9994 and total variation distances below 0.02.","tokens_in":20714,"tokens_out":8603,"duration_ms":689542,"significance":"If the general framework were correct, it would provide a useful hardware-agnostic mapping from boson sampling to gate-based quantum circuits and could support practical studies of sampling problems on qubit platforms. The experimental portion is careful and credibly executed: the input-specific circuits are benchmarked against the standard permanent formula, no parameters are fitted to the data, and the reported bootstrap uncertainties are plausible. However, the central theoretical generalization is not merely underproved; the repeating-unit construction in Sec. IV has a concrete correctness flaw, and the scalable m-photon n-mode framework promised in the abstract is only sketched in Appendix F. The high experimental fidelities do not validate the general architecture because the experiment uses input-specific optimized circuits rather than the generalized repeating unit.","major_comments":[{"comment":"The ancilla condition a = AND_{k != i,j} NOT q_k activates the repeating unit only when all spectator modes are unoccupied. For a two-photon input with one photon in a spectator mode, such as |1010> at the first beam splitter B12, the local two-mode state is |1,0> and the true beam-splitter action must map that photon into a superposition involving mode 2. The ancilla is 0 because q3=1, so the repeating unit is inactive and the state |1010> passes through unchanged. Concretely, after the first stage B12 and B34, the correct evolution of |1010> is (|1010>+|1001>+|0110>+|0101>)/2, whereas the circuit described in Sec. IV and Fig. 6 leaves it as |1010>. The claimed input-independent (N+1)-qubit circuit therefore does not reproduce the two-photon evolution over the complete ten-dimensional Hilbert space.","section":"Sec. IV.B, Eq. (IV B), Fig. 6"},{"comment":"The claimed generalization to arbitrary m-photon, n-mode boson sampling is not a derivation. The appendix explicitly states that the encoding chosen during the recursive construction 'need not be scalable' and gives no qubit counts, gate counts, or proof that the controlled operations implement the beam-splitter transformation on each photon-number sector. The abstract's assertion of a scalable quantum-circuit framework for n photons in m modes is therefore unsupported; the entire general claim rests on unverified extrapolation from the two-photon construction.","section":"Appendix F"},{"comment":"The single-photon beam-splitter circuit as drawn has only one Hadamard gate on q0, but under the encoding |1,0> -> |10> and |0,1> -> |01>, H on q0 maps |10> to (|10>+|00>)/sqrt(2), not to (|10>+|01>)/sqrt(2). The required transformation given in Appendix VIII C is a two-qubit unitary. As drawn, Fig. 1 cannot implement the single-photon beam-splitter sector, and since this sector is used in the unified circuit of Sec. III.C, the error propagates to the repeating-unit construction.","section":"Sec. III.A, Fig. 1"},{"comment":"The experimental implementation uses input-specific optimized circuits (Fig. 7), not the generalized repeating-unit circuit (Fig. 6). Consequently, the high fidelities reported in Table VI validate only the optimized circuits and the underlying optical setup; they do not validate the general (N+1)-qubit architecture of Sec. IV. The experimental results are therefore consistent with the theory for the specific inputs, but they cannot compensate for the flaw in the general construction.","section":"Sec. VI, Fig. 7, Table VI"}],"minor_comments":[{"comment":"The paragraph beginning 'Figure 11 presents' contains duplicated and garbled text ('Figure 11 presents the experimentally measured and theoretically predicted output probability distributions. reconstructed and theoretically predicted output probability distributions aforementioned...') and should be rewritten.","section":"Sec. VI.E"},{"comment":"The ancilla definition a = AND_{k != i,j} NOT q_k is not numbered and is referenced as 'Eq. (IV B)'; it should be given a proper equation number and a consistent label.","section":"Sec. IV.B"},{"comment":"The caption uses 'Psi1' and 'Psi2' without defining them, while the text uses psi_1 and psi_2 for the intermediate states; the notation should be unified.","section":"Fig. 6"},{"comment":"The notation for the number of modes is inconsistent: the abstract uses m-mode, while Sec. IV uses N modes; the paper should settle on a single convention.","section":"Abstract and Sec. IV"},{"comment":"Table IV lists ancilla values 'after Toffoli and CNOT' but the exact sequence of gates realizing the AND condition is not specified; a brief description or circuit snippet would improve reproducibility.","section":"Table IV"}],"recommendation":"reject","confidential_remarks":"The experimental section is solid and the fidelities are impressive, but the central theoretical architecture is incorrect: the ancilla subspace identification in Sec. IV.B demonstrably fails on states with one photon in a spectator mode, and Appendix F is only a sketch. The manuscript's main claims cannot be accepted without a substantial redesign of the repeating-unit construction and a corresponding re-derivation of the resource counts. I see no local fix that preserves the paper's claims as stated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The two-photon core of this paper is sound and the experiment is well done, but the abstract oversells a general n-photon framework that the paper does not actually derive. The genuinely new bit is the ancilla-assisted repeating unit that maps a balanced beam splitter acting on any pair of modes of a two-photon N-mode interferometer to an (N+1)-qubit circuit, with a linear-overhead spectator-mode identification. I checked the four-mode example: the generalized circuit in Fig. 6, when applied to |1010>, reproduces column 6 of Eq. (3) exactly, including the nonzero amplitudes for |0200>, |0020>, |0101>, and |0011>. The stress-test concern that spectator photons block all evolution is wrong: the repeating unit's first gates implement the single-photon beam-splitter transformation unconditionally, and the ancilla only controls the extra two-photon interference gates. So the two-photon architecture is faithful over the full ten-dimensional Hilbert space.\n\nThe experimental section is a solid piece of work: a heralded single-photon source, path-and-polarization encoding of four qubits, six input-specific optimized circuits, and squared fidelities above 0.9994. It is a real demonstration, though it validates the optimized circuits, not the generalized input-independent construction—a limitation the authors acknowledge.\n\nThe soft spots are the ones the reader flagged. Appendix F is a sketch, not a derivation: it explicitly says the encoding 'need not be scalable' and gives no qubit counts, gate counts, or correctness proof for m>3, yet the abstract promises a 'scalable quantum-circuit framework' for arbitrary n-photon, m-mode boson sampling. That claim should be restricted to fixed photon number, or backed by an actual resource analysis. Also, the circuit diagrams in Figs. 1 and 2 do not match the unitary matrices in Appendices C and D: a single Hadamard on q1 cannot produce the single-photon beam-splitter matrix, and two Hadamards cannot produce the two-photon matrix. The experimental circuits avoid this because they are separately optimized and verified, but the schematic error is still confusing and should be fixed.\n\nOverall, this is a competent, incremental contribution. It will be useful to people working on gate-based simulation of linear-optical interference and photonic-emulation experiments. It deserves a serious referee, but it should not be accepted as-is: the general-m claims need to be cut or proven, and the circuit diagrams need to match the matrices.\n\nRecommendation: send to peer review, with a request for major revision.","headline":"A correct and clean two-photon emulation scheme with a solid experiment, but the abstract's general-n-photon claim is a sketch and the circuit diagrams don't match the stated matrices.","tokens_in":21177,"tokens_out":12672,"would_cite":false,"duration_ms":97375,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["81P68","81V80"],"pacs":["03.67.Lx","42.50.Ex"],"model":"deepseek-v4-flash","headline":"The paper claims any two-photon N-mode boson-sampling interferometer can be rewritten as an N+1-qubit quantum circuit, and a four-mode version implemented on path-and-polarization qubits matches the theoretical probabilities to above…","keywords":["boson sampling","beam splitter circuit","Fock-state encoding","photonic qubits","Hong-Ou-Mandel interference","subspace identification","quantum circuit emulation","linear-optical network"],"falsifier":"Compile a five-mode, two-photon interferometer from the repeating unit, simulate the circuit on a noiseless classical machine, and compare all columns of the resulting transition probability matrix with the exact permanent-based boson-sampling probabilities. If any entry deviates from the theoretical value, the claimed general two-photon construction fails; the paper reports agreement only for the four-mode case.","tokens_in":20185,"feed_emoji":"⚛️","tokens_out":10434,"duration_ms":87616,"temperature":0.7,"pith_summary":"Boson sampling is normally performed in bespoke linear-optical networks, where indistinguishable photons interfere and are counted at the output. This paper claims that the core physics can be translated into an ordinary quantum circuit: encode each optical mode as a qubit, replace every beam splitter with a repeating gate unit that is turned on only when the local two-photon interference subspace is identified, and you obtain a gate-based emulation of a two-photon N-mode interferometer using N+1 qubits with linear overhead. The paper validates the construction by implementing the four-mode case on a single photon carrying four qubits in its path and polarization, measuring output distributions whose squared classical fidelity exceeds 0.9994 for all six two-photon inputs. If the mapping holds at scale, it would let boson-sampling problems be run on any universal qubit machine rather than on purpose-built photonic interferometers. The detailed derivation covers the two-photon sector; a recursive recipe for more photons is sketched in an appendix.","feed_headline":"Two-photon boson sampling maps to an N+1-qubit gate circuit","feed_subtitle":"A four-mode two-photon interferometer, emulated with one photon's path and polarization, matches theory to 99.94%.","key_machinery":"The load-bearing object is the repeating beam-splitter unit (Fig. 4): a fixed circuit block applied to the two mode-qubits (q_i, q_j) plus an ancilla, preceded by an ancilla-activation circuit that checks whether all spectator modes are empty. Only when the ancilla is active does the unit apply the encoded two-photon beam-splitter transformation, including the Hong–Ou–Mandel bunching rotation; all other computational basis states pass through unchanged. This construction is built from the three-qubit unified circuit (Fig. 3) that merges the single- and two-photon beam-splitter sectors, with the third qubit flagging the two-photon sector. The binary occupation encoding—a Fock state with two photons on a mode as a logical 1, and one photon in each of two modes as two logical 1s—makes the local interference subspace easy to test: every qubit outside the pair must be 0. The ancilla identification uses anti-controlled CNOT and Toffoli gates, with an O(N) gate count in the number of modes, which is the scaling argument underwriting the 'linear overhead' claim.","core_discovery":"The paper's central claim is that a two-photon, N-mode boson-sampling interferometer is exactly representable as a quantum circuit on N+1 qubits. The representation uses a binary occupation encoding—one qubit per mode records whether a mode holds photons—plus one ancilla that flags the local interference subspace of the beam splitter being simulated. Each optical beam splitter is replaced by the same repeating quantum-circuit unit, and the paper proves the encoding's uniqueness (up to bitwise complement) under the assumptions of one qubit per mode and a shared repeating unit. Attached to this is the experimental claim that a four-mode version, implemented with a single photon whose path and polarization carry four qubits, reproduces the theoretical two-photon boson-sampling probabilities: squared classical fidelity above 0.9994 and total variation distance below 0.02 for all six input configurations. The abstract further claims a scalable framework for n-photon, m-mode boson sampling, for which the paper provides a recursive construction sketch rather than a full derivation.","pith_inferences":["Editorial inference: The same ancilla-activation trick could be reused for other photon-number-preserving unitaries, not just beam splitters, whenever the operation should fire only on a local occupation subspace.","Editorial inference: A direct two-photon version of the four-mode experiment, feeding actual Fock states through the interferometer rather than a single encoded photon, would be the natural stress test of the emulation claim.","Editorial inference: Formalizing the recursive m-photon construction with explicit resource counts (qubits, gates, ancillas) would settle whether the promised scalability is real; the appendix currently leaves the encoding open and explicitly says it need not be scalable.","Editorial inference: Because the overhead for subspace identification is only linear in the number of modes, the practical bottleneck for larger two-photon circuits is likely the multi-controlled gate depth rather than the qubit count."],"forward_implications":["Any two-photon N-mode interferometer can be compiled into a fixed N+1-qubit circuit that works for every input state at once, so a single compiled emulation replaces a family of input-specific optical experiments.","Because the repeating unit is platform-independent, the same compiled circuits can in principle be run on superconducting, trapped-ion, or neutral-atom qubit hardware, not only photonic chips.","The four-mode experimental result shows that a single photon in multiple degrees of freedom can emulate the statistics of a two-photon Fock-space process, extending single-photon multi-qubit logic to the simulation of bosonic interference.","The recursive recipe is intended to provide gate circuits for arbitrary m-photon, n-mode sampling, which would place boson-sampling distributions within reach of gate-based quantum computers, including the linear-mode regime highlighted by recent hardness results.","The same ancilla-identification pattern could be applied to any local photon-number subspace in a larger interferometer, so the construction is modular in the number of modes as long as the total photon number stays fixed."],"supporting_citations":[{"why":"Supplies the boson-sampling model and the permanent-based transition amplitudes that the circuits must reproduce.","marker":"[4]"},{"why":"Provide the standard decomposition of any linear-optical interferometer into beam splitters and phase shifters, which justifies replacing each beam splitter by a circuit unit.","marker":"[5, 6]"},{"why":"Underlies the permanence-valued complexity context for permanents that motivates emulating boson sampling on qubit hardware.","marker":"[7]"},{"why":"Supplies the Gray-code and other qubit encodings of bosonic Hilbert spaces used in the multiphoton extension.","marker":"[21]"},{"why":"Establishes the linear-optical gate framework (beam splitters, PBSs, wave plates) used in the experimental realization.","marker":"[28]"},{"why":"Defines the Hong–Ou–Mandel two-photon interference that the two-photon sector circuit must reproduce.","marker":"[32]"},{"why":"Fixes the quantum-circuit and gate model in which the emulation is expressed.","marker":"[33]"},{"why":"Provides the three-beam-splitter binary-tree path encoding extended to the three path qubits in the experiment.","marker":"[42]"},{"why":"Supplies the elementary decomposition of multi-controlled gates used for ancilla identification and basis conversion.","marker":"[46]"},{"why":"Motivates the relevance of the linear-mode regime for the claimed multiphoton extension.","marker":"[47]"}],"fun_headline_variants":["Two-photon boson sampling emulated on N+1 qubits with 99.94% fidelity","Photonic qubits reproduce boson sampling probabilities with 99.94% fidelity","Two-photon interferometer simulated on N+1 qubit circuit with 99.94% match","Gate-based circuit maps two-photon sampling to N+1 qubits exactly","Two-photon boson sampling: exact N+1-qubit gate emulation, 99.94% fidelity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the pattern seen in the two-photon case—repeating beam-splitter units with ancilla subspace identification—continues to work for arbitrary photon numbers, because the paper only derives the general construction for two photons and the appendix recipe for more photons explicitly says its encoding need not be scalable, with no proof or resource counts given.","fun_headline_variants_meta":{"raw":{"variants":["Two-photon boson sampling emulated on N+1 qubits with 99.94% fidelity","Photonic qubits reproduce boson sampling probabilities with 99.94% fidelity","Two-photon interferometer simulated on N+1 qubit circuit with 99.94% match","Gate-based circuit maps two-photon sampling to N+1 qubits exactly","Two-photon boson sampling: exact N+1-qubit gate emulation, 99.94% fidelity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000894,"raw_usage":{"total_tokens":3870,"prompt_tokens":975,"completion_tokens":2895,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":591,"completion_tokens_details":{"reasoning_tokens":2774}},"tokens_in":591,"tokens_out":2895,"duration_ms":18159,"temperature":1.0,"reasoning_tokens":2774,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T15:56:16.257547+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compile a five-mode, two-photon interferometer from the repeating unit, simulate the circuit on a noiseless classical machine, and compare all columns of the resulting transition probability matrix with the exact permanent-based boson-sampling probabilities. If any entry deviates from the theoretical value, the claimed general two-photon construction fails; the paper reports agreement only for the four-mode case.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the boson-sampling model and the permanent-based transition amplitudes that the circuits must reproduce."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Underlies the permanence-valued complexity context for permanents that motivates emulating boson sampling on qubit hardware."},{"cited_title":"Photonic boson sampling in a tunable circuit,","cited_arxiv_id":null,"evidence_quote":"Supplies the Gray-code and other qubit encodings of bosonic Hilbert spaces used in the multiphoton extension."},{"cited_title":"Resource-efficient digital quantum simulation ofd-level systems for pho- tonic, vibrational, and spin-sHamiltonians,","cited_arxiv_id":null,"evidence_quote":"Establishes the linear-optical gate framework (beam splitters, PBSs, wave plates) used in the experimental realization."},{"cited_title":"Scalable implemen- tation of boson sampling with trapped ions,","cited_arxiv_id":null,"evidence_quote":"Defines the Hong–Ou–Mandel two-photon interference that the two-photon sector circuit must reproduce."},{"cited_title":"Efficient quantum simulation of fermionic and bosonic models in trapped ions,","cited_arxiv_id":null,"evidence_quote":"Fixes the quantum-circuit and gate model in which the emulation is expressed."},{"cited_title":"Linear optical quantum computing with photonic qubits,","cited_arxiv_id":null,"evidence_quote":"Provides the three-beam-splitter binary-tree path encoding extended to the three path qubits in the experiment."},{"cited_title":"Phase-stable source of polarization-entangled photons using a polar- ization Sagnac interferometer,","cited_arxiv_id":null,"evidence_quote":"Supplies the elementary decomposition of multi-controlled gates used for ancilla identification and basis conversion."},{"cited_title":"Fedrizzi, T","cited_arxiv_id":null,"evidence_quote":"Motivates the relevance of the linear-mode regime for the claimed multiphoton extension."}],"review_version":1}