{"id":"fb97bdd9-d9d2-49d1-bb12-02fd24987315","arxiv_id":"1908.07487","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Minimal modular extensions of super-Tannakian categories are classified using fermionic actions and group cohomology, yielding explicit counts for examples like Z/6Z and Z/4Z.","lead":"This mathematics paper classifies, in cohomology terms, the minimal modular extensions of super-Tannakian categories, which are symmetry categories with both bosonic and fermionic degrees of freedom. The result gives a concrete method to count such extensions for finite super-groups, with implications for topological phases of matter.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 6.14's converse is a proof gap: fermionic action does not obviously imply the associated modules lie in Pic(B,f), and the classification rests on this.","rationale":"The reader's verdict is CONDITIONAL, and this stress-test agrees that the main classification is not fully established. The strongest claim, Theorem 7.12, is a cohomological parametrization of the preimage of D, but it depends on Theorem 6.14, whose proof is a sketch. The gap is not a known counterexample; the worked examples agree with known results, and much of the framework comes from ENO10 and GVR17. However, the missing details are exactly the ones that connect the module-category side (Pic(B,f)) to the autoequivalence side (fermionic actions). Without an explicit proof that theta_{D_g} = g* and that Pic(B,f) is a 2-subgroup, the correspondence between braided (G~,z)-crossed extensions and 2-homomorphisms is an assertion rather than a theorem. I therefore keep the verdict CONDITIONAL: the classification should be accepted only after the converse of Theorem 6.14 and the 2-equivalence in Proposition 6.13 are supplied.","tokens_in":20612,"tokens_out":20947,"duration_ms":212562,"concrete_test":"Re-derive the missing converse of Theorem 6.14 in full: using the equivalence D_g = Fun_B(D_g,D_g) and equations (3)-(4), prove theta_{D_g} is isomorphic to g* as braided autoequivalences and that this identification sends the condition g*(f) = f to theta_{D_g}(f) = f. Then check that the tensorator of the 2-homomorphism is compatible with the monoidal structure of Pic(B,f). If this derivation cannot be completed, Theorem 7.12 is unsupported; as a numerical cross-check, enumerate 2-homomorphisms Z/2 -> Pic(C,f) for the eight pointed dimension-4 modular categories and compare the total with the 32 minimal modular extensions of Rep(Z/4Z,[2]) claimed in Example 7.17.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is assembled from Proposition 7.10, Corollary 7.11 and Theorem 7.12, and the load-bearing step is Theorem 6.14: braided (G~,z)-crossed extensions of (B,f) are identified with 2-homomorphisms G -> Pic(B,f) whose truncation is a fermionic action. The proof of the converse direction is incomplete. Given a braided (G~,z)-crossed extension D, ENO10 Theorem 7.12 yields a 2-homomorphism G -> Pic(B), but one must prove that every component D_g lies in the full subcategory Pic(B,f), i.e. that theta_{D_g}(f) is isomorphic to f. The proof asserts \"if D_g in Pic(B,f) then g* is a fermionic functor\", which is the reverse of the needed implication. What is required is that the fermionic action hypothesis g* in Autbr(B,f) implies theta_{D_g}(f) = f, via an explicit identification of theta_{D_g} with g* using the alpha-induction formulas (3)-(4). Proposition 6.13 only checks one containment on objects and does not establish that theta restricts to a 2-equivalence of categorical groups. If theta_{D_g} differs from g* by a twist that does not preserve f, or if the monoidal coherence fails, the bijection of Corollary 7.11 and the triple parametrization of Theorem 7.12 would overcount or undercount the preimage under D.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a cohomological classification of minimal modular extensions of super-Tannakian categories, i.e. of the group Mext(Rep(~G,z)). The main tool is a fermionic analogue of the ENO10 correspondence between braided G-crossed extensions of a braided fusion category B and 2-group homomorphisms G -> Pic(B). The paper states this analogue as Theorem 6.14, then uses it in Section 7 to describe the preimage of the map D: Mext(Rep(~G,z)) -> Mext(SVec) in terms of triples (rho, mu, phi) with rho: G -> Autbr(C,f), mu in a torsor over the kernel of r*: H^2_rho(G,K0(C)) -> H^2_rho(G,K0(SVec)), and phi in a torsor over H^3(G,C^×), subject to vanishing of obstructions O3(rho,alpha) and O4(rho,mu). Concrete consequences are drawn in Theorem 7.15 and Example 7.17, giving orders for Mext(Rep(Z/mZ x Z/2Z)) and Mext(Rep(Z/4Z)).","tokens_in":20840,"tokens_out":9416,"duration_ms":85116,"significance":"If the classification is correct, it provides an explicit cohomological parametrization of minimal modular extensions of super-Tannakian categories, a problem left open since LKW16a. The paper also gives a fermionic version of a central result of ENO10, extends the author's earlier work with Galindo, and derives concrete numerical predictions (e.g. 16m for trivial super-groups and 32 for Z/4Z) that can be checked against known classifications. The strategy of using equivariantization/de-equivariantization and the Picard 2-group is natural and appropriate. However, the proofs of the central statements are sketches: the converse direction of Theorem 6.14 contains a gap that is load-bearing for the whole parametrization, and Theorem 7.12 is stated with a derivation that essentially restates the theorem. The numerical examples in Section 7.3 rely on phrases such as 'a similar analysis' and are not fully justified. These issues affect the central claim and require substantial revision.","major_comments":[{"comment":"The converse direction of the claimed bijection is not proved. Starting from a braided (~G,z)-crossed extension D of (B,f), the proof invokes ENO10 Theorem 7.12 to obtain a 2-homomorphism ~~rho: G -> Pic(B), and then asserts that 'the G-action ... is fermionic' and 'if D_g in Pic(B,f) then g* is a fermionic functor'. The second statement is the converse of what is needed: the hypothesis gives g* in Autbr(B,f), and one must prove that D_g lies in the full subcategory Pic(B,f), i.e. that theta_{D_g}(f) is isomorphic to f. This requires identifying theta_{D_g} with g* through the alpha-induction formulas (3)-(4) and showing that the module-category condition 'the module functors -⊗f and f⊗- are isomorphic' is equivalent to theta_M(f) ≅ f. The proof also does not explicitly verify that the braided G-crossed extension produced from a 2-homomorphism into Pic(B,f) has faithful G-grading with trivial component exactly B. Since Corollary 7.11 and Theorem 7.12 are built on this bijection, the central classification is not established without this step.","section":null},{"comment":"The parametrization of the preimage by triples (rho,mu,phi) is the paper's central result, but the proof is largely a restatement. The text asserts that 'any 2-homomorphism associated by truncation to rho can be parametrized by an element in H^2_rho(G, K0(C)) × H^3(G,C^×)' and that the fermionic condition is 'equivalent' to O3(rho,alpha)=0 and mu in Ker(r*), without demonstrating the needed torsor structures. In particular, one must show that the torsor of liftings of rho to an action restricts to a torsor over the kernel of r*: H^2_rho(G,K0(C)) -> H^2_rho(G,K0(SVec)), that each such lifting can be extended to a 2-homomorphism taking values in Pic(C,f), and that this extension is compatible with the choice of phi in a torsor over H^3(G,C^×). The vanishing of O4(rho,mu) is stated, but O4 is defined for a fully specified bosonic action; it is not explained how O4 is evaluated for a fermionic action or why the condition is independent of the choices made. As written, the converse direction of the proof does not establish the classification.","section":null},{"comment":"The explicit orders 16m and 32 rest on unproved assertions. In Theorem 7.15(b), the proof says 'the unique group homomorphism Z/mZ -> Z/2Z × Z/2Z is the trivial homomorphism', but the target of rho is Autbr(Vec^{ω0,c0}_{Z/2Z×Z/2Z}, f), not the group of invertible objects; the reader must infer the automorphism group has order two and that all homomorphisms from Z/mZ to it are trivial. In Example 7.17(c), the claim that 'a similar analysis shows that every pointed fusion category with fusion rules given by Z/2Z × Z/2Z is in the image of D' is not a proof, and item (b) 'Z/4Z is not a trivial super-group, so no Ising category can be in the image of D' does not follow from Corollary 7.8 alone, which only implies D is not surjective. The numerical conclusions of the paper depend on these facts, so they need to be proved or explicitly justified.","section":null},{"comment":"The definition of Pic(B,f) uses the condition theta_M(f)=f, but theta_M is defined only up to natural isomorphism, so the condition should be theta_M(f) ≅ f. The statement that this is equivalent to the module functors -⊗f and f⊗- being isomorphic autoequivalences of M is given without proof. Proposition 6.13 then asserts an equivalence between Autbr(B,f) and Pic(B,f), but the proof only checks essential surjectivity of the restriction of theta; full faithfulness and the monoidal structure of the restricted equivalence are not addressed. This equivalence is one of the ingredients of Theorem 6.14, and the missing details are needed for the bijection to be rigorous.","section":null}],"minor_comments":[{"comment":"The paragraph after the heading 'Equivariantization' begins with the fragment 'processes of equivariantization and de-equivariantization are some of the main tools...' and appears to be missing a sentence; the text should be cleaned up.","section":null},{"comment":"Remark 7.13 refers to 'Proposition 7.12', but the statement being discussed is Theorem 7.12.","section":null},{"comment":"The notation 'Ker(r*: H^2_rho(G, K0(C) -> H^2_rho(G, K0(SVec))' is missing a closing parenthesis after the domain of r*; moreover, the proof switches between K0(SVec) and Z/2Z without explaining the identification.","section":null},{"comment":"The paper relies heavily on results from the author's companion preprint [GVR17] (Theorems 3.8, 4.2, and 7.6, and the fermionic obstruction O3) without proofs. If [GVR17] is not yet published, the author should either include the statements with proofs or indicate where they are available in a peer-reviewed form.","section":null},{"comment":"The count of '4 such triples' would be clearer if the size of H^2(Z/2Z, Z/2Z × Z/2Z) and the effect of the condition r*(mu) nontrivial were stated explicitly.","section":null}],"recommendation":"major_revision","confidential_remarks":"The manuscript depends heavily on the companion paper [GVR17] by the same author, including the fermionic action framework, the obstruction O3, and Theorem 7.6 on existence of minimal modular extensions. The editor may wish to confirm the publication status of [GVR17] before evaluating the independence of the present results. There is also a notable gap in Theorem 6.14's converse that is load-bearing for the classification; I would recommend requesting a revised version with a complete proof of that step and of the torsor parametrization in Theorem 7.12, rather than rejecting on the basis of the current sketches."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a credible and useful attempt at Müger's problem, but the central theorem has a real proof gap that needs to be closed before the classification is trustworthy.\n\nWhat is actually new: Theorem 6.14 states a fermionic analogue of ENO10's classification of braided G-crossed extensions, and Theorem 7.12 parametrizes the preimage under D: Mext(Rep(G~,z)) -> Mext(SVec) by triples (rho, mu, phi) with cohomological obstructions. That is a genuine extension of the ENO10/LKW machinery to super-Tannakian categories, and it is the first concrete approach to the case left open by LKW16a. The examples for Z/6Z and Z/4Z matching known counts (48 and 32) are a useful sanity check, not a proof, but they do indicate the framework is not wildly off.\n\nWhere the soft spots are: The proof of Theorem 6.14's converse is not just terse, it has the implication backwards. Given a braided (G~,z)-crossed extension, you need to show the associated 2-homomorphism G -> Pic(B) lands in the subcategory Pic(B,f), i.e. that theta_{D_g}(f)=f for each g. The text instead asserts that if D_g is in Pic(B,f) then g* is fermionic, which is the reverse direction. The missing step is an explicit identification of theta_{D_g} with the action functor g*, via the alpha-induction formulas (3)-(4). Proposition 6.13 has the same flavour: it proves one containment but claims an equivalence. None of this is likely fatal — the statement is probably true — but the classification in Theorem 7.12 and the kernel computation in Corollary 7.14 depend on exactly this bijection, so the gap matters. The proof of Theorem 7.12 itself is also a sketch: the parametrization by triples is asserted rather than derived, and Example 7.17 leans on 'a similar analysis' for a non-trivial count. And the paper leans heavily on [GVR17] for the obstruction theory; that is legitimate if [GVR17] is solid, but a referee should check those black boxes rather than take them on faith.\n\nThe reasoning style is serious, the examples match known results, and the classification, once fixed, would be a real tool for fermionic topological orders. This deserves a serious referee, with the request to fill the gap in 6.14 and expand the proofs of 7.12 and 7.17. I would not cite it as a finished result until those steps are written out.","headline":"A plausible cohomological classification of minimal modular extensions of super-Tannakian categories that rests on a key fermionic Picard correspondence whose proof is currently sketched, not finished.","tokens_in":21456,"tokens_out":3238,"would_cite":false,"duration_ms":31611,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["18D10","20J06"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper classifies minimal modular extensions of super-Tannakian categories by triples of group-cohomology data, subject to two vanishing obstructions.","keywords":["super-Tannakian categories","minimal modular extensions","fermionic actions","braided G-crossed extensions","spin-braided fusion categories","Picard 2-group","group cohomology","equivariantization"],"falsifier":"A concrete test: search for a braided $(\\widetilde G,z)$-crossed extension of an Ising spin-braided category by a non-trivial super-group; the paper's Proposition 5.1 says only the trivial super-group acts fermionically on Ising categories, so an explicit example would disprove the obstruction theory, while the vanishing conditions of Theorem 7.12 should rule it out.","tokens_in":20337,"feed_emoji":"🧮","tokens_out":16480,"duration_ms":139247,"temperature":0.7,"pith_summary":"Minimal modular extensions are the smallest modular categories containing a given braided fusion category, and the problem of finding them for symmetric categories was solved in the bosonic (Tannakian) case but left open for the super-Tannakian case. This paper claims that for a finite super-group $(\\widetilde G,z)$, every minimal modular extension of the super-Tannakian category $\\operatorname{Rep}(\\widetilde G,z)$ is encoded by a braided $(\\widetilde G,z)$-crossed extension of a minimal modular extension of super-vector spaces, and that such extensions are classified by triples $(\\rho,\\mu,\\phi)$: a group homomorphism into the spin-braided autoequivalence group, a class in a restricted $H^2$-torsor, and a class in an $H^3$-torsor, subject to the vanishing of two obstructions $O_3$ and $O_4$. If correct, this turns the classification into finite group cohomology. The paper works out the resulting counts for examples: $16m$ extensions for $\\mathbb{Z}/m\\mathbb{Z}\\times \\mathbb{Z}/2\\mathbb{Z}$ with $m$ odd, $48$ for $\\mathbb{Z}/6\\mathbb{Z}$, and $32$ for $\\mathbb{Z}/4\\mathbb{Z}$.","feed_headline":"Cohomology triples classify super-Tannakian modular extensions","feed_subtitle":"Every minimal extension is a triple: a group map and classes in H^2 and H^3, with two obstructions required to vanish.","key_machinery":"The central machinery is the spin-braided Picard 2-group $\\operatorname{Pic}(B,f)$: the full subcategory of the Picard 2-group of a braided fusion category $B$ whose objects are invertible module categories $M$ whose induced braided autoequivalence $\\theta_M$ fixes the distinguished fermion $f$. Proposition 6.13 identifies the truncation of $\\operatorname{Pic}(B,f)$ with $\\operatorname{Aut}^{\\mathrm{br}}_{\\otimes}(B,f)$, the group of spin-braided autoequivalences. The argument funnels every extension through this object: by Theorem 6.14 a braided $(\\widetilde G,z)$-crossed extension is a 2-group homomorphism $G\\to \\operatorname{Pic}(B,f)$, and truncating that homomorphism and then lifting the data back to cohomology produces the triple $(\\rho,\\mu,\\phi)$. The two obstructions carry the existence burden: $O_3(\\rho,\\alpha)$ decides when $\\rho$ lifts to a fermionic action, and $O_4(\\rho,\\mu)$ decides when that fermionic action lifts to a 2-group homomorphism into the Picard group.","core_discovery":"Let $(\\widetilde G,z)$ be a finite super-group, write $G=\\widetilde G/\\langle z\\rangle$, and let $\\alpha\\in H^2(G,\\mathbb{Z}/2\\mathbb{Z})$ be the class determined by the extension $1\\to \\langle z\\rangle\\to \\widetilde G\\to G\\to 1$. The central claim, Theorem 7.12, fixes a minimal modular extension $C$ of $\\operatorname{SVec}$ and describes the pre-image of $C$ under the group homomorphism $D:\\operatorname{Mext}(\\operatorname{Rep}(\\widetilde G,z))\\to \\operatorname{Mext}(\\operatorname{SVec})$. Every element of that pre-image is parametrized by a triple $(\\rho,\\mu,\\phi)$, where $\\rho:G\\to \\operatorname{Aut}^{\\mathrm{br}}_{\\otimes}(C,f)$ is a group homomorphism, $\\mu$ lies in a torsor over the kernel of $r_*:H^2_\\rho(G,K_0(C))\\to H^2_\\rho(G,K_0(\\operatorname{SVec}))$, and $\\phi$ lies in a torsor over $H^3(G,\\mathbb{C}^\\times)$; the two obstructions $O_3(\\rho,\\alpha)$ and $O_4(\\rho,\\mu)$ must vanish. The route to the theorem is the fermionic analogue of the standard bosonic correspondence: Theorem 6.14 states that braided $(\\widetilde G,z)$-crossed extensions of a spin-braided fusion category $(B,f)$ are in bijection with 2-group homomorphisms $G\\to \\operatorname{Pic}(B,f)$ whose truncation is a fermionic action. On the paper's view, a minimal modular extension of $\\operatorname{Rep}(\\widetilde G,z)$ is exactly such a fermionic crossed extension, so the cohomological triple follows from the 2-group classification.","pith_inferences":["The paper leaves implicit that the same triple classification should apply to any slightly degenerate braided fusion category whose symmetric center is $\\operatorname{SVec}$, not only to super-Tannakian categories; if Theorem 6.14 holds, the obstruction pair $(O_3,O_4)$ is a general counting tool for fermionic modular extensions.","The worked examples suggest a parity dichotomy: for supergroups with non-zero class $\\alpha\\in H^2(G,\\mathbb{Z}/2\\mathbb{Z})$, Ising-type extensions are excluded from the image of $D$ while pointed extensions always appear; testing this on supergroups with even-order $G$, such as $\\mathbb{Z}/8\\mathbb{Z}$, would be a direct extension of the paper's computations.","Because $\\operatorname{Mext}(\\operatorname{Rep}(\\widetilde G,z))$ is an abelian group and $D$ is a homomorphism, the triples of Theorem 7.12 should carry an explicit composition law; spelling it out would present the kernel of $D$ as a group extension of $H^3(G,\\mathbb{C}^\\times)$ by the pointed part.","Since all data in the theorem are finite group-cohomology sets, the classification is algorithmically checkable for any finite supergroup; independent counts for small groups such as $\\mathbb{Z}/8\\mathbb{Z}$ or $\\mathbb{Z}/9\\mathbb{Z}$ would test the formula beyond the examples in the paper."],"forward_implications":["For a fixed minimal modular extension $C$ of $\\operatorname{SVec}$, the fiber $D^{-1}(C)$ is a cohomology set built from $H^2$ and $H^3$, so once $\\rho$ is chosen, constructing an extension is a finite calculation of two torsors and two obstructions.","For the supergroup $\\mathbb{Z}/4\\mathbb{Z}$ with distinguished element $[2]$, there are exactly $32$ minimal modular extensions of $\\operatorname{Rep}(\\mathbb{Z}/4\\mathbb{Z},[2])$; the kernel of $D$ has order $4$ and the image of $D$ consists of the pointed modular extensions of $\\operatorname{SVec}$.","For odd $m$, the trivial supergroup $\\mathbb{Z}/m\\mathbb{Z}\\times \\mathbb{Z}/2\\mathbb{Z}$ has exactly $16m$ minimal modular extensions.","The homomorphism $D$ is surjective if and only if the supergroup is trivial; for non-trivial supergroups, minimal modular extensions lie over the subgroup of $\\operatorname{Mext}(\\operatorname{SVec})$ that admits a fermionic crossed extension.","Each admissible triple constructs a braided crossed extension whose equivariantization gives a minimal modular extension, so the classification is constructive rather than merely counting."],"supporting_citations":[{"why":"Supplies the bosonic bijection between braided G-crossed extensions and 2-group homomorphisms G to Pic(B), which Theorem 6.14 adapts to the fermionic setting.","marker":"[ENO10]"},{"why":"Defines fermionic actions, the obstruction O3(rho,alpha), the homomorphism D, and the reduction of minimal modular extensions to crossed extensions; the paper's whole problem setup comes from here.","marker":"[GVR17]"},{"why":"Establishes that minimal modular extensions of a symmetric category form an abelian group Mext and classifies the Tannakian case, giving the target of D and the counting framework.","marker":"[LKW16a]"},{"why":"Provides the equivariantization and de-equivariantization bijections and the classification of dimension-four modular categories used in the examples.","marker":"[DGNO10]"},{"why":"Gives the obstruction O3(rho) and the torsor description of liftings over H^2_rho(G,K0(C)), a key input to the triple parametrization.","marker":"[Gal11, Theorem 5.5]"},{"why":"Supplies the H^4-obstruction O4(rho,mu) deciding when a categorical 2-group homomorphism can be lifted to Pic(B).","marker":"[CGPW16, Proposition 9]"},{"why":"Deligne's classification identifies symmetric fusion categories as Rep(G) or Rep(tilde G,z), fixing the super-Tannakian categories as the object of study.","marker":"[Del02]"}],"fun_headline_variants":["Fermionic extensions pinned by H^2 and H^3 triples","Minimal modular extensions: a cohomology triple recipe","Two obstructions vanish: triple parametrizes extensions","Spin-braided categories: minimal modular extensions","Cohomology triple classifies minimal fermionic extensions"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The classification rests on the claim that every crossed extension with a fermionic action is represented by an invertible module category preserving the distinguished fermion; the proof of that direction in the paper leaves the equivalence between these two conditions unproved.","fun_headline_variants_meta":{"raw":{"variants":["Fermionic extensions pinned by H^2 and H^3 triples","Minimal modular extensions: a cohomology triple recipe","Two obstructions vanish: triple parametrizes extensions","Spin-braided categories: minimal modular extensions","Cohomology triple classifies minimal fermionic extensions"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000988,"raw_usage":{"total_tokens":4210,"prompt_tokens":984,"completion_tokens":3226,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":600,"completion_tokens_details":{"reasoning_tokens":3144}},"tokens_in":600,"tokens_out":3226,"duration_ms":23252,"temperature":1.0,"reasoning_tokens":3144,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:18:25.468523+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete test: search for a braided $(\\widetilde G,z)$-crossed extension of an Ising spin-braided category by a non-trivial super-group; the paper's Proposition 5.1 says only the trivial super-group acts fermionically on Ising categories, so an explicit example would disprove the obstruction theory, while the vanishing conditions of Theorem 7.12 should rule it out.","supporting_citations":[],"review_version":1}