{"id":"616845e7-ab50-4634-9e20-4528c9781552","arxiv_id":"2508.13856","paper_version":1,"verdict":"UNVERDICTED","confidence":"UNKNOWN","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"This submission pairs an abstract about fair multi-stage assignment (C-Balance, DC-Balance, NP-hardness) with a body that is a different paper about quantum-optical duality, leaving the abstract's claims unsupported.","lead":"The abstract announces algorithms with envy and cost guarantees for fair assignment on multi-stage graphs, plus an NP-hardness proof. The body text is an unrelated quantum optics paper, so none of the announced results are supported anywhere in this submission.","discovery_kind":"unclear","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Abstract's load-bearing claims (C-Balance bounds, NP-hardness, CoF, experiments) are absent from the body, which is an unrelated quantum-optics paper; the central results cannot be checked in this manuscript.","rationale":"The reader's verdict is UNVERDICTED, and my stress-test identifies the same dominant concern: the manuscript does not contain the technical content needed to evaluate the central claims. The weakest technical premise identified by the reader — that envy is bounded by 2M independent of K — is indeed fragile, but it is secondary here because the algorithms and proofs are not present at all. The optics body is not bad science; it is a different paper. Treating all manuscript text as in-scope, the mismatch is the decisive fact. I do not manufacture an additional technical objection, because none could be assessed from the available text. The concrete test — a source-level search for the absent terms — would settle the existence question directly. Since the reader already chose UNVERDICTED for this reason, my recommendation leaves that verdict unchanged.","tokens_in":22582,"tokens_out":2305,"duration_ms":25565,"concrete_test":"Fetch the arXiv source for 2508.13856, open the main TeX file, and search the body text for the terms 'C-Balance', 'DC-Balance', 'envy', 'multi-stage', and 'cost of fairness'; also compare the title, abstract, and author list in the source against the submission metadata. If no occurrence of these terms appears in the body and the body title matches arXiv:2508.13855v3, the abstract's results are unsupported and the manuscript is unverifiable as submitted. A secondary check would be to search arXiv for any separately posted version of the assignment paper and compare its theorems against the abstract's claims.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Every load-bearing assertion in the abstract — NP-hardness of envy-minimizing assignment on multi-stage graphs, C-Balance's 2M envy bound and its tightness, CoF ≤ 2, DC-Balance convergence to envy arbitrarily close to 2M, and orders-of-magnitude speedup over an ILP — depends on definitions, theorems, algorithms, and experiments that never appear. The body text is a quantum-optics paper (arXiv:2508.13855v3) proving a nonlinear-linear duality for nondegenerate PDCs; it contains no multi-stage graph, no assignment problem, no envy measure, no C-Balance/DC-Balance, and no ILP comparison. The optics body is internally coherent and honestly states its own limitations (no single-mode squeezing or real optical cavities; postselection amplitude calculation remains hard), but none of that bears on the assignment claims. The submission-level premise — that this manuscript contains the technical content promised by the abstract — is false. Consequently there is no way to audit the two-agent envy bound's dependence on M rather than K, the tightness instance, or the CoF bound. This is an absence-of-evidence concern, not a disagreement with consensus.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The submission has the title and abstract of a paper on fair assignment in multi-stage graphs: it claims NP-hardness of envy-minimizing assignment, a C-Balance algorithm with a tight 2M envy bound for two agents, cost-of-fairness bound 2, a DC-Balance extension, and experimental speedups over an ILP. The full text, however, is a quantum-optics paper titled 'Nonlinear-linear duality for multipath quantum interference' with its own author list, introduction, theory, appendices, and references, and it contains none of the assignment problem, algorithms, definitions, theorems, proofs, or experiments announced in the abstract. The abstract's central claims are therefore uncheckable in this manuscript.","tokens_in":22670,"tokens_out":2803,"duration_ms":31154,"significance":"If the claimed results were stated and proved, they would constitute a substantive contribution: a first approximation framework with worst-case envy and cost guarantees for fair path assignment on multi-stage graphs, including an NP-hardness result and an algorithmic comparison with ILP. However, the manuscript as submitted provides none of these results. The optics body may be a legitimate and internally coherent contribution, with proofs in appendices and an honest statement of its own limitations, but it is not the paper described by the abstract. Consequently, the significance of the claimed results cannot be evaluated, and the submission-level premise — that this manuscript contains the promised technical content — is false.","major_comments":[{"comment":"None of the central objects or results appear in the body. The abstract asserts: (1) NP-hardness of envy-minimizing assignment on multi-stage graphs; (2) C-Balance guarantees envy at most 2M for two agents, with tightness; (3) cost-of-fairness at most 2; (4) DC-Balance converges to envy arbitrarily close to 2M; (5) orders-of-magnitude speedup over a suitably formulated ILP. Sections I–IV and Appendices A–F contain no multi-stage graphs, no assignment problem, no envy measure, no C-Balance or DC-Balance algorithms, and no ILP. These are not presentation issues; they are the entire claimed content of the paper, and their absence makes the abstract's assertions impossible to audit.","section":"Abstract vs. Full text"},{"comment":"The two-agent envy bound is claimed to be 2M, with M the maximum edge weight, and no stated dependence on the number of stages K. Since an agent's path cost is a sum of edge weights over many stages, pairwise cost disparity could in principle grow with K even when M is fixed. The manuscript provides no definition of envy, no algorithm, and no proof, so this fragile premise cannot be assessed. No equation or theorem beyond the abstract can be cited to check whether the bound depends only on M.","section":"Abstract (envy bound)"},{"comment":"The abstract claims 'we experimentally show that our algorithm runs several orders of magnitude faster than a suitably formulated ILP.' The full text has no experimental section, no ILP formulation, no benchmarks, and no runtime comparisons. Moreover, the Data Availability statement in the body says 'No data were created or analyzed in this study.' This directly contradicts the abstract's empirical claim and makes the reported speedup unverifiable.","section":"Data Availability / Experiments"}],"minor_comments":[{"comment":"The title of the submission, 'The Multi-Stage Assignment Problem: A Fairness Perspective,' does not match the body's title, 'Nonlinear-linear duality for multipath quantum interference.' The body's arXiv header shows 2508.13855v3, not 2508.13856. This mismatch must be resolved; as submitted, the document is not self-consistent.","section":"Title and metadata"},{"comment":"The body's reference list is entirely on quantum optics and contains no citations to multi-stage assignment, fair division, envy minimization, or computational complexity related to the abstract's claims. This reinforces that the body is a different paper.","section":"References"}],"recommendation":"reject","confidential_remarks":"The submission appears to be a file or metadata mix-up: the body is the quantum-optics paper arXiv:2508.13855v3, while the abstract and title describe an unrelated multi-agent assignment paper. No revision could add the missing theorems, algorithms, and experiments without effectively rewriting the entire manuscript. If the intended assignment paper exists, it should be resubmitted as a clean, self-contained manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis submission is two different papers under one cover. The title and abstract describe an algorithmic game theory paper on envy-minimizing assignments in multi-stage graphs, with claimed NP-hardness, a C-Balance algorithm with envy bound 2M and cost-of-fairness bound 2, a DC-Balance extension, and ILP speedups. The body text is \"Nonlinear-linear duality for multipath quantum interference\" by Zheng et al. (arXiv:2508.13855v3). There is no multi-stage graph, no assignment problem, no envy, no C-Balance, no DC-Balance, no ILP. Nothing in the abstract's technical claims can be checked in this manuscript.\n\nWhat is actually present is the quantum-optics paper, and that paper looks coherent: it proves a generalized duality for nondegenerate PDC networks, uses Gaussian-state and transfer-matrix arguments, includes lemmas and proofs, and explicitly states its own limits (no single-mode squeezing, no real cavities, postselection amplitudes remain hard). So the body is not junk; it just has nothing to do with the abstract.\n\nThe soft spot is the entire submission-level premise. Every load-bearing assertion depends on definitions and proofs that never appear. The envy bound's dependence on M instead of path length K, the tightness instance, the CoF ratio, the convergence claim—all unverifiable. The reader's concern is exactly right: this is an absence-of-evidence problem, not a disagreement with consensus.\n\nCould the abstract's claims be true? Possibly. Multi-stage assignment with fairness constraints is a reasonable area, and the claimed bounds are specific enough that they might be derivable. But this manuscript is not a vehicle for evaluating them. Sending this to peer review would waste referee time unless the correct body is restored.\n\nIf the authors accidentally uploaded the wrong file, they should resubmit with the matching assignment paper. If this is not an accident, it's a serious submission-integrity problem. Either way, the current version should not advance.\n\nFor you: don't cite the abstract, don't chase the results, and flag this if you see it in the wild. The optics paper itself, if submitted separately under its own title, would deserve a serious referee—but that's not the paper we were asked to judge.\n\nRecommendation: desk reject the current version; do not send the assignment claims to peer review until they actually exist in a manuscript.","headline":"Abstract advertises a fair-assignment paper, but the body is an unrelated quantum-optics paper; none of the claimed results exist in this manuscript.","tokens_in":23350,"tokens_out":2501,"would_cite":false,"duration_ms":22744,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This submission claims envy-minimizing assignment on multi-stage graphs is NP-hard and gives a 2M envy bound, yet the body text is an unrelated quantum-optics paper.","keywords":["multi-stage graphs","fair assignment","envy minimization","cost of fairness","NP-hardness","C-Balance","DC-Balance","submission mismatch"],"falsifier":"Read the submitted full text in search of C-Balance, the NP-hardness reduction, and theorem statements; none appears. To test the mathematical bound directly, implement the algorithm described in the abstract on two-agent instances of 3 or more stages with maximum edge weight 1 and search for envy exceeding 2; the first such instance would refute the 2M guarantee even before the missing proofs are supplied.","tokens_in":22289,"feed_emoji":"⚠️","tokens_out":5049,"duration_ms":51544,"temperature":0.7,"pith_summary":"This submission's abstract proposes a framework for fairly assigning node-disjoint paths to agents on a multi-stage graph: it asserts that minimizing envy is NP-hard, that the C-Balance algorithm guarantees envy at most 2M for two agents (with a tight example), that the cost of fairness is at most 2, and that DC-Balance iteratively extends this to n agents while running far faster than an ILP. The full text deposited with the submission, however, is a different manuscript on nonlinear-linear duality in quantum optics: none of the algorithms, definitions, or proofs from the abstract appears in the body. As submitted, then, the paper's stated contribution cannot be checked, and the only verifiable fact is the mismatch between the abstract and the full text. A corrected manuscript would be needed to evaluate the claimed results.","feed_headline":"Abstract promises fair assignment; body is quantum optics","feed_subtitle":"The claimed NP-hardness result and 2M envy bound appear nowhere in the submitted text.","key_machinery":"The stated machinery is the C-Balance algorithm: for two agents it supposedly equalizes assigned path costs so that envy—the maximum cost gap between agents—is at most 2M, with a matching lower-bound example; DC-Balance repeatedly invokes C-Balance to handle n agents and is claimed to converge to envy arbitrarily close to 2M. The underlying object is a multi-stage graph, a sequence of weighted bipartite graphs between adjacent layers in which each agent receives a node-disjoint path from the first stage to the last. None of this machinery is present in the submitted full text, which concerns beamsplitter replacements for parametric down-conversion.","core_discovery":"The abstract asserts that an assignment minimizing total path cost on a multi-stage graph—nodes partitioned into K stages with weighted bipartite edges between consecutive stages, agents routed along node-disjoint paths from first to last stage—can be highly envious, and that finding an envy-minimizing assignment is NP-hard. It further asserts C-Balance, a two-agent balancing algorithm, bounds pairwise envy by 2M where M is the maximum edge weight, that this bound is tight, that the cost of the fair assignment is at most twice the minimum-cost assignment, and that DC-Balance makes iterative calls to C-Balance to bring n-agent envy arbitrarily close to 2M. The body text provided with this sub","pith_inferences":["The abstract–body mismatch reads as a packaging error in the submission rather than a deliberate mathematical claim; the assignment results should be evaluated only against a complete, corrected manuscript.","If the 2M bound is later proven, the most natural stress test is whether it degrades with the number of stages K; a reader should look for a counterexample with many small edges whose sum creates envy beyond 2M.","The quantum-optics body has its own standalone content, but it cannot be used as support for, or evidence about, the multi-stage assignment problem."],"forward_implications":["If the abstract's claims hold, fair assignment on multi-stage graphs becomes tractable in a practical sense: a polynomial algorithm with constant worst-case envy in the largest edge weight and a cost-of-fairness factor of 2.","The NP-hardness result would justify the approximation approach, ruling out exact envy minimization in general.","The tight 2M example would show the analysis cannot be improved within the same algorithmic strategy.","The claimed orders-of-magnitude speedup over ILP would make the algorithm usable in large staged routing settings."],"supporting_citations":[],"fun_headline_variants":["Abstract promises fair assignment; body is quantum optics","NP-hardness and 2M envy bounds missing from paper","Multi-stage assignment: claims in abstract, proofs absent","Fair assignment results claimed, not supported in body","Abstract oversells: no NP-hardness proof in text"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that a pairwise cost gap between two agents is controlled by the largest single edge weight M, even though each agent's cost is a sum over many edges; the even weaker premise, at submission level, is that the algorithms and proofs described in the abstract exist in this manuscript at all.","fun_headline_variants_meta":{"raw":{"variants":["Abstract promises fair assignment; body is quantum optics","NP-hardness and 2M envy bounds missing from paper","Multi-stage assignment: claims in abstract, proofs absent","Fair assignment results claimed, not supported in body","Abstract oversells: no NP-hardness proof in text"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000679,"raw_usage":{"total_tokens":2960,"prompt_tokens":819,"completion_tokens":2141,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":563,"completion_tokens_details":{"reasoning_tokens":2064}},"tokens_in":563,"tokens_out":2141,"duration_ms":17261,"temperature":1.0,"reasoning_tokens":2064,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T18:52:27.235163+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Read the submitted full text in search of C-Balance, the NP-hardness reduction, and theorem statements; none appears. To test the mathematical bound directly, implement the algorithm described in the abstract on two-agent instances of 3 or more stages with maximum edge weight 1 and search for envy exceeding 2; the first such instance would refute the 2M guarantee even before the missing proofs are supplied.","supporting_citations":[],"review_version":1}