{"id":"3669e5b5-616f-41af-9acb-87da950eb228","arxiv_id":"2607.23039","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":2.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper reproduces, with more computational detail, the known Molien series and 21-generator invariant ring for two-qubit mixed states under local unitary equivalence.","lead":"This paper re-derives the known structure of polynomial invariants of two-qubit mixed states under local unitary transformations, including an explicit contour-integral computation of the Molien series and a graphical construction of 21 ring generators. It is an exposition of results from King et al. (2007) aimed at making the derivation more accessible.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Graph enumeration in Theorem 4.4 is asserted by inspection; a missed connected graph would break the claimed 21-generator result.","rationale":"I read the paper in good faith and checked the Molien-series derivation separately: the Weyl measure normalization, the eigenvalue factors of the integrand, and the low-degree expansion (n0=1, n1=1, n2=4) are consistent. The central claim therefore rests on the constructive part, Theorem 4.4. The proof's binary is: (i) all connected graphs are enumerated, and (ii) every non-retained candidate reduces to the 21 generators. Both are supported by diagrams and asserted reductions rather than by an explicit exhaustive method. This is precisely the reader's weakest assumption, and it is load-bearing: the Molien series cannot detect a missing generator. The concern does not amount to rejecting the theorem—it is a known result, and the evidence presented is largely consistent—but it does mean the paper's self-described 'complete' derivation is not fully substantiated. A computational enumeration would settle the issue. The minor typo in part (v) (⟨a|C|a⟩ instead of ⟨a|C|b⟩) reinforces the need for a machine-checkable step but is not itself the main concern. Verdict remains CONDITIONAL, so no change from the reader's verdict.","tokens_in":26883,"tokens_out":11975,"duration_ms":114407,"concrete_test":"Implement an exhaustive graph-generation program for cases (24), (26), and (31) using the building blocks a, b, C, ε_A, ε_B, enforcing the chain-length bound from Proposition 4.1 and using graph isomorphism to avoid duplicates. Translate every generated connected graph into a polynomial in a, b, C, then compute its normal form modulo the ring generated by the 21 listed invariants in a computer algebra system (e.g., Macaulay2 or Singular). If every such polynomial reduces to zero, the enumeration is complete; if even one graph gives a nonzero normal form, Theorem 4.4's generator set is incomplete as stated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 4.4's proof reduces the generator question to enumerating all connected graphs in cases (24), (26), and (31). The text supports this only by 'By enumerating all the connected graphs, it can be known that...' and by diagrams (Figures 7–11). No algorithm, computer algebra check, or rigorous counting argument is given. Because the Molien series only fixes the dimensions of homogeneous invariant spaces and cannot certify generation, a missed graph would be an invariant not expressible by the listed 21, invalidating the central constructive claim. This concern is sharpened by the quick dismissals of cases (29)/(30) and by the reduction of all case-(31) graphs to Q1/Q2 for 'arbitrary A-type and B-type vectors'; both steps rely on unstated assumptions about which C-C chains and cross-product identities exhaust all possibilities. The claim is also not independently supported by formal verification. The paper would be fully convincing if the enumeration were replaced or supplemented by an exhaustive computational check.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper revisits the known characterization, due to King, Welsh, and Jarvis, of the ring of local unitary polynomial invariants of two-qubit mixed states. It has two main parts: first, a detailed derivation of the Molien series for the action of SU(2)×SU(2) on the 16-dimensional space of density matrices, carried out by reducing to the maximal torus, applying Weyl integration, and evaluating the resulting double contour integral by residues; second, a graphical tensor method that constructs candidate invariants from Bloch-vector and correlation-matrix building blocks, reduces the candidate list using algebraic identities, and claims that the remaining 21 invariants generate the full invariant ring. The paper is explicitly framed as an exposition that fills in computational and pedagogical gaps in the original treatment. The final Molien series is stated in Theorem 3.5, and the 21-generator claim is stated in Theorem 4.4.","tokens_in":27109,"tokens_out":6181,"duration_ms":62973,"significance":"If the derivation is accepted, the paper provides a valuable self-contained exposition of a structurally important result: the Molien series is computed in unusual detail, and the expansion reproduces the known low-degree counts n0=1, n1=1, n2=4, which is a useful sanity check. The paper also gives an explicit, organized list of the 21 generating invariants and derives several nontrivial reduction identities. The main result is not new, but the expository value for quantum information and invariant theory audiences is real. However, the constructive part of the proof, especially the exhaustiveness of the graphical enumeration, is not yet presented at the standard of rigor that would make the paper a reliable reference.","major_comments":[{"comment":"The proof of Theorem 4.4 rests on the claim that the connected graphs listed in Figures 7–11 exhaust all possibilities for cases (24), (26), and (31). The only support is the sentence after Figure 8: 'By enumerating all the connected graphs, it can be known that ...'. No algorithm, no counting argument, and no computer algebra check is supplied. This is load-bearing: the Molien series fixes only the dimensions of homogeneous invariant spaces and cannot by itself certify that the 21 listed polynomials generate the invariant ring. A missed connected graph could produce an invariant not expressible in terms of the 21 generators, invalidating the constructive claim. Because Proposition 4.1 bounds C-chains by at most four C-nodes, the enumeration is finite and could be made machine-checkable. I request either a rigorous combinatorial enumeration with an explicit counting bound, or a reproduci","section":"§4.2 (vii)–(ix), Appendix C"},{"comment":"Even accepting the diagrams as exhaustive, the reduction in case (31) is not itemized. The text states that 'from these diagrams' all candidate invariants except U1, P1, P2 reduce to Q1 or Q2, but no per-graph reduction is given, and the 'arbitrary A-type and B-type vectors' in Q1 and Q2 are not defined with respect to which subgraphs they represent. A reader cannot verify that each of the numerous unlabeled graphs in Figures 9–11 is either zero, reducible to a product of lower-degree invariants, or expressible through the listed reductions. Please provide either the explicit polynomial identity for every non-generating graph or a script that performs these reductions symbolically.","section":"§4.2 (ix), Figures 9–11"},{"comment":"The treatment of case (24) is similarly terse. Figure 7 displays many graphs, but only one ('W4') is labeled as a surviving candidate; the others are dismissed with the comment that they 'can either be generated by other invariants or are identically zero.' No indication is given of which graph uses which identity. Since case (24) is one of the three cases on which the exhaustiveness of the 21-generator set depends, this omission weakens the proof of Theorem 4.4.","section":"§4.2 (vii), Figure 7"}],"minor_comments":[{"comment":"The text says that situation (16) can only generate '⟨a|C|a⟩ and ⟨a|CC^TC|a⟩', but Table 3 and Figure 3 identify the surviving invariants as K6 = ⟨a|C|b⟩ and U2 = ⟨a|CC^TC|b⟩. This appears to be a typo and should be corrected to avoid confusion.","section":"§4.2 (v)"},{"comment":"There is a typo in the concluding remarks: 'Gerdtet al' should be 'Gerdt et al.'","section":"Section 5"},{"comment":"The notation 'bC' and similar expressions is used before its definition as the transpose-adjugate of C. A short explicit sentence defining these symbols in one place, preferably near Table 1, would improve readability.","section":"§4.2, equations after (4.1)"},{"comment":"The captions of Figures 7–11 do not explain the color/red marking convention or the status of the unlabeled graphs. Adding a note such as 'red labels indicate surviving invariants; unlabeled graphs are excluded by one of the stated identities' would make the figures much easier to check.","section":"Appendix C"},{"comment":"The reductions of Q1 and Q2 use 'Lemmas B1 and B2' from reference [10] without restating them. For a paper whose goal is accessibility, stating these two lemmas explicitly in the text or an appendix would be helpful.","section":"§4.2 (ix)"}],"recommendation":"major_revision","confidential_remarks":"The paper is an exposition of a known result, so the threshold for novelty is appropriately lower, but the manuscript currently presents an incomplete proof of its central constructive theorem. The missing piece—exhaustive verification of the graph enumeration—is well within scope and can be fixed with a computational supplement or a careful counting argument. I would encourage the editor to request that fix rather than reject, because the Molien-series part is careful and the paper fills a genuine pedagogical gap."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: this is an exposition, not new science. It revisits King–Welsh–Jarvis's 2007 characterization of the two-qubit LU-invariant ring and fills in the Molien-series computation step by step via residue calculus. That part is well done. The closed form checks out in the low-degree expansion (1, 1, 4 as expected), and the contour derivation is explicit enough to follow.\n\nWhat's genuinely useful: the paper makes the Molien integral tractable, with all poles and residues spelled out, and it lays out the graphical tensor method for constructing the 21 generators. The attribution is honest—the theorems are credited to King et al. and Grassl et al. If a student or a researcher in quantum information wants to see how that q^10−q^8−... series emerges from SU(2)×SU(2), this is a good place to look.\n\nThe soft spot is exactly where you'd expect: Theorem 4.4's proof that the 21 listed invariants generate the ring depends on the enumeration of connected graphs in cases (24), (26), and (31), and that enumeration is asserted by inspection ('By enumerating all the connected graphs, it can be known that...'). There's no algorithm, no computer algebra script, no counting argument. The Molien series only fixes dimensions; it can't certify generation. So if a connected graph was missed, the 21 generators could be incomplete. The authors also reduce case (31) to Q1 and Q2 'for arbitrary A- and B-type vectors' with some unstated assumptions about cross-product identities and C-C chains. This is a real gap, but it's an addressable one: a short exhaustive check (even a brute-force graph enumeration script) would close it.\n\nMinor complaint: the abstract says 'filling in all previously omitted computational steps,' which overstates. Some simplifications are still just asserted ('After simplifying, one obtains...'), and the completeness claim in Section 4.2 is exactly the kind of step that was omitted in the original. The paper would be strengthened by toning that down.\n\nOverall: the main theorems are known and correctly attributed; the new contribution is exposition, and the exposition is good where it is complete. The graph-enumeration gap is the one thing that keeps this from being fully convincing. For a journal like J. Phys. A or similar, I would send it to a referee who knows invariant theory and can assess whether the enumeration is plausible; with a supplementary computational check, I'd be comfortable accepting. For a reading group, it's a decent teaching resource, but not essential.","headline":"A readable, honest re-derivation of King et al.'s two-qubit invariant ring results, with the Molien series done properly; the generator proof still has an enumeration-by-inspection gap that should be closed before publication.","tokens_in":27600,"tokens_out":1979,"would_cite":false,"duration_ms":18946,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["13A50","13H10","81P40"],"pacs":[],"model":"deepseek-v4-flash","headline":"21 polynomials generate all two-qubit local-unitary invariants.","keywords":["two-qubit systems","local unitary invariants","invariant ring","Molien series","graphical tensor method","integrity basis","entanglement classification","Cohen-Macaulay"],"falsifier":"Use computer algebra to compute the Hilbert series of the ring generated by the 21 listed invariants and compare it coefficient-by-coefficient with the Molien series M_G(q) above; the first degree at which they differ would reveal a missed generator, while agreement to high degree would confirm the enumeration.","tokens_in":26750,"feed_emoji":"⚛️","tokens_out":7086,"duration_ms":65889,"temperature":0.7,"pith_summary":"This paper aims to establish, with all computational steps written out, that the ring of polynomial invariants of a two-qubit density matrix under local unitary transformations is generated by exactly 21 explicitly listed polynomials, and that its Molien series—the generating function counting independent invariants degree by degree—is a specific rational function. The Molien series is derived from scratch by contour integration over the maximal torus of SU(2)×SU(2), after reducing GL(2)×GL(2) to its compact subgroup. The generator set is built by a graphical method: every invariant corresponds to a fully contracted graph built from the Bloch vectors a, b, the correlation matrix C, and the antisymmetric tensors of SO(3); enumerating connected graphs and applying reduction identities prunes 31 candidate cases down to 21 irreducible generators. If correct, this gives a complete, explicit algebraic description of local unitary equivalence for two-qubit mixed states, which is the basic classification tool for two-qubit entanglement.","feed_headline":"21 polynomials generate all two-qubit local-unitary invariants","feed_subtitle":"Complete generator list plus closed-form Molien series pin down the invariant ring's structure.","key_machinery":"The argument rests on two pieces. First, the Molien integral: after reducing GL(2)×GL(2) to SU(2)×SU(2) (scalar matrices act trivially, and SU(2) is Zariski dense in SL(2)), the representation on the 16-dimensional vectorized density matrix is diagonalized on the maximal torus diag(e^{iα/2}, e^{-iα/2})⊗diag(e^{-iα/2}, e^{iα/2})⊗diag(e^{iβ/2}, e^{-iβ/2})⊗diag(e^{-iβ/2}, e^{iβ/2}), turning the Haar integral into a double contour integral in z = e^{iα}, w = e^{iβ}. Repeated residue computations at the poles inside the unit circle yield the closed-form series. Second, the graphical tensor method: in the Bloch picture the action is SO(3)×SO(3), and all invariants are complete contractions built f","core_discovery":"The paper's central claim is that the invariant ring for two-qubit mixed states under G = SU(2)×SU(2) (equivalently GL(2)×GL(2)) has Molien series M_G(q) = (q^10 − q^8 − q^7 + 2q^6 + 2q^5 + 2q^4 − q^3 − q^2 + 1) / ((1−q)^10 (1+q)^6 (1+q^2)^2 (1+q+q^2)^3), and is generated as a C-algebra by the 21 invariants K1,...,K9, X1, X2, U1, U2, V1,...,V4, W1,...,W4. The paper supplies two things earlier treatments left opaque: a residue-calculus proof of the Molien series with every pole computed, and a systematic graph enumeration explaining why exactly these 21 are needed and why the other candidates reduce to combinations of them. The result is framed as a rigorous, self-contained exposition of a 20","pith_inferences":["If the enumeration is complete, the same graphical calculus can be turned into a computer program: generate all connected graphs up to the bounded chain length and verify the Hilbert series coefficient-by-coefficient against the Molien series, testing the generator claim without trusting hand-drawn figures.","The gap between the 21 generators and the 18 invariants known to separate LU-equivalence classes suggests a two-tier structure—a small separation set plus extra generators needed to close the ring; making this gap explicit could help choose minimal measurement sets for entanglement detection.","For qubit-qutrit systems, the paper's own remarks imply the contraction-based approach stalls at the absence of an SO(8) image; a likely workaround is to work directly with SU(3) invariant tensors rather than SO(8), at the cost of many more graph types.","A numerical sampler could test the 21 generators empirically: generate random two-qubit states, evaluate a high-degree invariant known from the Molien-series dimension, and check that the 21 generators span it; disagreement would localize a gap in the enumeration."],"forward_implications":["The 21 polynomials form an integrity basis: every polynomial local-unitary invariant of a two-qubit state is a polynomial in these 21, so any LU-invariant quantity is expressible, in principle, in terms of them.","The Molien series gives the full graded dimension sequence of the invariant ring; from the denominator factors one can read off the degrees of a homogeneous system of parameters and the secondary invariants in a Hironaka decomposition.","The explicit reduction identities—chain shortening and cross-product decomposition—show that only finitely many graph patterns need be checked, making the generator proof checkable by hand.","The graphical strategy is specific to the two-qubit case: the paper argues it cannot be directly ported to qubit-qutrit systems because no surjective homomorphism SU(d)→SO(d^2−1) exists for d>2.","The candidate enumeration includes two extra invariants that were missing from earlier enumerations, and the paper shows they reduce to combinations of the 21, confirming that the 21 suffice."],"fun_headline_variants":["Two-qubit invariant ring pinned down by 21 explicit generators","Graphical proof finds 21 invariants that generate the two-qubit ring","All two-qubit invariants: 21 polynomials from a closed-form Molien series","Complete two-qubit invariant ring: 21 generators, rigorous Molien series"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The whole generator proof rests on the assertion—made by visual inspection of the figures in Appendix C—that the connected graphs drawn for cases (24), (26), and (31) are all the connected graphs that can occur; if any connected graph was missed, the 21 polynomials might generate a proper subring of the invariant ring even though the Molien series is correct.","fun_headline_variants_meta":{"raw":{"variants":["Two-qubit invariant ring pinned down by 21 explicit generators","Graphical proof finds 21 invariants that generate the two-qubit ring","All two-qubit invariants: 21 polynomials from a closed-form Molien series","Complete two-qubit invariant ring: 21 generators, rigorous Molien series"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000255,"raw_usage":{"total_tokens":1439,"prompt_tokens":807,"completion_tokens":632,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":551,"completion_tokens_details":{"reasoning_tokens":550}},"tokens_in":551,"tokens_out":632,"duration_ms":6579,"temperature":1.0,"reasoning_tokens":550,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T03:48:02.415761+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Use computer algebra to compute the Hilbert series of the ring generated by the 21 listed invariants and compare it coefficient-by-coefficient with the Molien series M_G(q) above; the first degree at which they differ would reveal a missed generator, while agreement to high degree would confirm the enumeration.","supporting_citations":[],"review_version":1}