{"id":"3c9ae33b-0410-4b2a-a0f1-8e6d0903fb1c","arxiv_id":"2505.00587","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For any C*-correspondence with a generalized gauge action by a locally compact Hausdorff group, the Cuntz-Pimsner algebra of the reduced crossed product is canonically isomorphic to the reduced crossed product of the Cuntz-Pimsner algebra.","lead":"This paper proves the Hao-Ng isomorphism theorem for reduced crossed products in full generality, a major open problem in operator algebras since 2008. It shows that the Cuntz-Pimsner algebra construction and the reduced crossed product construction commute for every locally compact Hausdorff group.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 4.3 assumes, without proof or citation, the existence of a faithful non-degenerate representation of C*_e(A) with the unique extension property; without such a representation the proof has no starting point.","rationale":"The reader's conditional verdict is justified, but the specific concern raised by the reader—the invariance principle in Proposition 4.2—is not the most load-bearing issue. The invariance principle is elementary: for any CP extension φ of (π∘α_g)|A, the map φ∘α_g^{-1} is a CP extension of π|A, so uniqueness for π forces φ = π∘α_g; the non-unital adaptation is routine because α_g is a completely isometric automorphism. The more serious gap is the unsupported existence assertion at the start of Theorem 4.3: the proof needs a single representation of C*_e(A) that is simultaneously faithful and has UEP. Known results guarantee a faithful direct sum of boundary representations, but the UEP of a direct sum is not automatic, and the Bilich–Dor-On counterexample shows how UEP can be lost in the passage from fibers to a direct integral. If this existence assumption fails, the subsequent argument cannot even be initialized, because σ would not be injective on C*_e(A)⋊G and the final equality would only describe a quotient. Therefore the paper should either supply a proof or reference for the existence of a faithful non-degenerate UEP representation, or modify the strategy to avoid needing a single faithful UEP representation. Since this is a precise, fixable gap rather than a demonstrated falsehood, the conditional verdict from the reader remains appropriate.","tokens_in":13514,"tokens_out":34998,"duration_ms":360219,"concrete_test":"Test the existence of a faithful UEP representation on a known non-hyperrigid operator algebra, e.g., the tensor algebra of the C*-correspondence from the Bilich–Dor-On counterexample (arXiv:2404.05018). Concretely, form the direct sum of all boundary representations of its C*-envelope and attempt to construct a UCP extension that agrees with it on A but differs on a self-adjoint element; if such an extension exists, no faithful UEP representation is obtained this way, and the proof of Theorem 4.3 would require a new argument. Alternatively, locate an explicit existence theorem in [23] or Arveson [4] and state it in the paper; if such a theorem is found, the gap closes.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The proof of Theorem 4.3 opens by fixing a non-degenerate injective ∗-representation π : C*_e(A) → B(H) that has the unique extension property with respect to A. For an arbitrary approximately unital operator algebra, no theorem is cited for the existence of such a π, and it is not a formal consequence of the results quoted in §2.1. The standard noncommutative Choquet theorem provides a faithful direct sum of boundary representations, but a direct sum of representations with UEP need not itself have UEP; indeed, the recent counterexample to Arveson's hyperrigidity conjecture (Bilich–Dor-On, arXiv:2404.05018, §3) shows that UEP can fail for direct integrals even when it holds fiberwise. Without a faithful UEP representation π, the representation σ = πα ⋊ (id⊗λ) on C*_e(A)⋊G has a nonzero kernel, so the equality C*_e(σ(A⋊G)) = σ(C*_e(A)⋊G) identifies only a quotient of C*_e(A)⋊G with the C*-envelope, not the full algebra. The final inference that σ is completely isometric covers A⋊G, but not the generated C*-algebra. The invariance principle in Proposition 4.2 is, by contrast, defensible: conjugating any CP extension with α_g^{-1} reduces uniqueness for π∘α_g to uniqueness for π, so the citation to [23] is adequate. The load-bearing gap is the existence of a faithful UEP representation, not the invariance principle.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves the reduced Hao-Ng isomorphism theorem for C*-correspondences over locally compact Hausdorff groups. The main results are Theorem A, asserting that the reduced crossed product functor commutes with the C*-envelope for operator algebras with a self-adjoint contractive approximate identity, and Theorem B, asserting that for a generalized gauge action on a non-degenerate C*-correspondence X, the Cuntz-Pimsner algebra of the reduced crossed product correspondence is canonically isomorphic to the reduced crossed product of the Cuntz-Pimsner algebra. The proof strategy follows Katsoulis-Ramsey: Theorem A is established first via operator-valued Maharam lifting (based on Hamana's Fubini tensor products), and then Theorem B is deduced from Theorem A together with established results on tensor algebras. The paper is carefully written and pays close attention to non-separable Hilbert space issues.","tokens_in":13807,"tokens_out":34181,"duration_ms":360063,"significance":"If the proof is correct, this resolves the reduced Hao-Ng isomorphism problem in full generality for locally compact Hausdorff groups, a long-standing open problem. The paper introduces a new technical tool---an operator-valued Maharam lifting theorem---to overcome the failure of the unique extension property to be preserved under direct sums, which was highlighted by the recent counterexample to Arveson's hyperrigidity conjecture. The writing is detailed and the reliance on imported results, such as Hamana's Fubini tensor products and Ionescu-Tulcea's lifting theorem, is explicit. The manuscript also honestly discusses limitations and the failure of the analogous statement for full crossed products. The main concern is that two load-bearing steps in the proof are cited to unpublished or unstated results without enough justification.","major_comments":[{"comment":"The proof begins by fixing a non-degenerate injective *-representation π : C*_e(A) → B(H) that has the unique extension property with respect to A, but no proof or citation is given for the existence of such a representation when A is a non-unital approximately unital operator algebra. This is load-bearing because Proposition 4.2 and the subsequent conclusion θ∘πα = πα rely on the unique extension property of π. The existence is in fact a standard consequence of the Dritschel-McCullough maximal dilation theorem applied to the unitization A∼ and then restricting to C*_e(A), but the manuscript should state and prove this as a lemma or give a precise reference for the non-unital setting.","section":"Section 4, Theorem 4.3"},{"comment":"The proof invokes an 'invariance principle' cited to the unpublished note [3, Proposition 3.1] and to [23, Proposition 2.4] to conclude that π∘α_g has the unique extension property with respect to A. Since [3] is unpublished and the exact non-unital hypotheses are not spelled out, this step is not verifiable from the manuscript alone. The authors should either prove the invariance principle in a few lines (by conjugating any extension with α_g^{-1}) or state the precise version with all hypotheses, because the uniqueness of πα depends on it.","section":"Section 4, Proposition 4.2"}],"minor_comments":[{"comment":"There is a typo in the abstract: 'W e prove' should be 'We prove'.","section":"Abstract"},{"comment":"The existence of representations with the unique extension property is discussed for unital operator algebras via Dritschel-McCullough, but the non-unital counterpart used in Theorem 4.3 is not stated. Please add an explicit statement for the approximately unital case.","section":"Section 2.1"},{"comment":"The phrase 'Let D be the C*-subalgebra of A generated by a self-adjoint contractive approximate unit of A' is potentially confusing because A is not self-adjoint. It would be clearer to say that D is the C*-subalgebra of the ambient C*-algebra generated by the elements of a self-adjoint contractive approximate identity, which is contained in A.","section":"Section 4, Theorem 4.3"},{"comment":"The identification Cb(G;B(H)) ∼= B(H)⊗Cb(G) is used without comment. For non-separable H, B(H)⊗Cb(G) is understood via Hamana's Fubini tensor product and is an operator system, so a brief justification of this identification would help the reader.","section":"Section 4, Proposition 4.2"},{"comment":"The reliance on the unpublished note [3] for the invariance principle is risky; consider including a short proof in the paper or using [23] as the primary reference.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper appears to be mathematically sound in its main ideas, and the missing justifications for the existence of a faithful UEP representation and for the invariance principle are likely repairable. The result is significant and fits the journal well. I would encourage the editor to request a revision that fills these gaps; the current manuscript, as written, does not give the reader enough to verify two load-bearing steps."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, the punchline: this paper settles the reduced Hao-Ng isomorphism for every locally compact Hausdorff group, going beyond the amenable, exact-discrete, and hyperrigid cases that were previously known. That is a real advance, and the main structural idea — proving that the reduced crossed product commutes with the C*-envelope for approximately unital operator algebras — is the right route. The operator-valued Maharam lifting built on Hamana's Fubini tensor products is a nice, careful piece of work; the authors are right to stress that B(H)⊗M∞(F) is only an operator system in the non-separable setting, and they handle that delicately.\n\nWhere the paper has a soft spot is the very opening of Theorem 4.3. It fixes a non-degenerate injective *-representation π of C*_e(A) with the unique extension property with respect to A. The paper does not prove or cite a theorem guaranteeing such a representation exists. This is not an idle technicality: as soon as the boundary representations are infinite, the usual faithful object is their direct sum, and the unique extension property is not preserved under direct sums or direct integrals — the Bilich–Dor-On counterexample cited by the authors makes this explicit. If no faithful UEP representation exists, the argument in Theorem 4.3 identifies C*_e(A⋊_{α,r}G) with a quotient of C*_e(A)⋊_{α,r}G, not the full algebra. That would leave the main theorem unproved.\n\nThe invariance principle, by contrast, is fine — conjugating a CP extension by α_g^{-1} reduces uniqueness for π∘α_g to uniqueness for π, so the citation to [3] and [23] is adequate. The abstract's promise of 'refined W*-dynamical covers' is stronger than what the body actually uses; that is a minor mismatch. The dependence on previous work such as [8] and [13] is legitimate and not circular; the main technical work is new.\n\nIf the authors can supply a proof, or a precise reference, that the C*-envelope of an approximately unital operator algebra admits a faithful non-degenerate UEP representation, the proof goes through. My suspicion is that such a theorem may hold for operator algebras even where it fails for general operator systems, but it is not stated and not obvious. That gap is exactly what a referee should press on.\n\nOverall: this is a serious paper, the main result is very likely correct, and the machinery introduced will be used elsewhere. It deserves rigorous peer review, but only after the existence question is settled in the revision. I'd bring it to the reading group, and I'd cite it once it's in final form.","headline":"A genuinely new and substantial result — the reduced Hao-Ng isomorphism in full locally compact generality — with a proof that is mostly rigorous but has one unaddressed existence gap that a serious referee should press on.","tokens_in":14366,"tokens_out":7654,"would_cite":true,"duration_ms":73228,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["46K50","47L55","46L55","46L05"],"pacs":[],"model":"deepseek-v4-flash","headline":"The reduced Hao-Ng isomorphism is proved for every locally compact Hausdorff group, by commuting the C*-envelope with reduced crossed products.","keywords":["Hao-Ng isomorphism","reduced crossed product","Cuntz-Pimsner algebra","C*-envelope","non-self-adjoint operator algebra","unique extension property","operator-valued Maharam lifting","Fubini tensor product"],"falsifier":"Construct a C*-dynamical system $(B,G,\\alpha)$, an $\\alpha$-invariant operator subalgebra $A$ generating $B$, and a representation $\\pi$ with the unique extension property with respect to $A$ such that $\\pi \\circ \\alpha_g$ fails that property for some $g \\in G$; then the uniqueness step of Proposition 4.2 collapses. The decisive test is whether $C^*_e(A \\rtimes_{\\alpha,r} G) \\cong C^*_e(A) \\rtimes_{\\alpha,r} G$ still holds in such an example, since a failure there would disprove the paper's central claim.","tokens_in":13276,"feed_emoji":"🔄","tokens_out":13896,"duration_ms":122931,"temperature":0.7,"pith_summary":"Every locally compact Hausdorff group action on a C*-correspondence now commutes with the Cuntz-Pimsner construction: the Cuntz-Pimsner algebra of the reduced crossed-product correspondence is canonically isomorphic to the reduced crossed product of the Cuntz-Pimsner algebra. This is the reduced Hao-Ng isomorphism, proved here for all non-degenerate C*-correspondences and all locally compact Hausdorff groups, after earlier results had covered amenable groups, discrete exact groups, and hyperrigid cases. A generalized gauge action is an action that preserves the coefficient algebra and the correspondence itself inside the algebra they generate. The proof's engine is a commutation theorem for the C*-envelope, the smallest C*-algebra generated by an operator algebra through which every completely isometric representation factors, and this theorem is applied to the tensor algebra of the correspondence.","feed_headline":"Reduced Hao-Ng isomorphism proved for all locally compact groups","feed_subtitle":"Cuntz-Pimsner algebras now commute with reduced crossed products for any locally compact Hausdorff group action.","key_machinery":"The load-bearing mechanism is a refined unique extension property. A representation $\\pi$ of the generated C*-algebra $B$ has the unique extension property with respect to $A$ when $\\pi|_A$ has exactly one completely contractive completely positive extension to $B$. Proposition 4.2 proves that for such a $\\pi$, the lifted representation $\\pi_\\alpha$ with range in $B(H) \\otimes L^\\infty(G)$ is uniquely determined on all of $B$ among completely contractive completely positive extensions of $\\pi_\\alpha|_A$ with range in that von Neumann algebra. The proof needs pointwise evaluation of operator-valued functions on $G$, and to make this work without separability assumptions the paper builds an operator-valued Maharam lifting theorem: the quotient map $B(H) \\otimes M^\\infty(\\mathcal{F}) \\to B(H) \\otimes L^\\infty(G)$ has a unital completely positive right inverse that acts as the identity on bounded continuous functions. Pointwise evaluation then pushes uniqueness at each group element up to uniqueness of the whole extension, and this is what makes the canonical map from $C^*_e(A) \\rtimes_{\\alpha,r} G$ to $C^*_e(A \\rtimes_{\\alpha,r} G)$ injective.","core_discovery":"The central claim is Theorem B: for a non-degenerate C*-correspondence $X$ over a C*-algebra $B$ and a generalized gauge action $\\alpha$ of a locally compact Hausdorff group $G$, there is a canonical $*$-isomorphism $\\mathcal{O}_{X \\rtimes_{\\alpha,r} G} \\cong \\mathcal{O}_X \\rtimes_{\\alpha,r} G$. The route is Theorem A: if $A$ is an operator algebra with a self-adjoint contractive approximate identity and $G$ acts on $A$, then the canonical map $C^*_e(A) \\rtimes_{\\alpha,r} G \\to C^*_e(A \\rtimes_{\\alpha,r} G)$ is an isomorphism. Since the Cuntz-Pimsner algebra of $X$ is the C*-envelope of the tensor algebra of $X$, applying Theorem A to that tensor algebra yields Theorem B.","pith_inferences":["The operator-valued Maharam lifting theorem is likely reusable beyond this proof: any argument that needs to evaluate completely positive maps on $L^\\infty(G)$ pointwise after lifting to bounded measurable functions could use the same right inverse, for instance in coaction or non-separable product-system problems.","A boundary test for Theorem A is to drop the self-adjointness of the approximate identity. If the unique extension property still behaves well under group twists for operator algebras with merely contractive approximate identities, the commutation theorem would extend; if not, the self-adjointness assumption is carrying real weight.","Earlier applications of the Hao-Ng isomorphism, such as AF-embeddability questions, were limited to the cases where the isomorphism was known; with this theorem those applications now have the full locally compact Hausdorff statement available, although the paper does not enumerate them."],"forward_implications":["For a non-degenerate C*-correspondence $X$ over $B$ and a generalized gauge action of any locally compact Hausdorff group $G$, the canonical isomorphism $\\mathcal{O}_{X \\rtimes_{\\alpha,r} G} \\cong \\mathcal{O}_X \\rtimes_{\\alpha,r} G$ holds.","For every operator algebra $A$ with a self-adjoint contractive approximate identity and action of a locally compact Hausdorff group, the C*-envelope commutes with the reduced crossed product: $C^*_e(A \\rtimes_{\\alpha,r} G) \\cong C^*_e(A) \\rtimes_{\\alpha,r} G$.","Since graph C*-algebras arise as Cuntz-Pimsner algebras of graph correspondences, reduced crossed products of graph C*-algebras by arbitrary locally compact Hausdorff group actions now fall under the theorem, with no amenability, exactness, or hyperrigidity assumption.","The paper's strategy does not by itself settle the full Hao-Ng isomorphism: the analogous commutation for full crossed products fails for arbitrary operator algebras and is left open for tensor algebras."],"supporting_citations":[{"why":"Supplies the invariance principle for the unique extension property that Proposition 4.2 invokes but does not prove.","marker":"[3]"},{"why":"Defines the unique extension property for non-unital operator algebras and contains the invariance discussion cited in the twisting step.","marker":"[23]"},{"why":"Provides the minimal S-projection argument used to realize the injective envelope inside a given injective range.","marker":"[25]"},{"why":"Constructs the Fubini tensor products that make the operator-valued Maharam lifting theorem work.","marker":"[27]"},{"why":"Identifies the Cuntz-Pimsner algebra of a correspondence with the C*-envelope of its tensor algebra.","marker":"[37]"},{"why":"Supplies the operator-algebra crossed product framework and the independence of the reduced crossed product from the chosen embedding.","marker":"[38]"},{"why":"Provides the hyperrigid version of the C*-envelope commutation and the proof strategy that this paper extends beyond hyperrigidity.","marker":"[39]"},{"why":"Gives the equivariant lifting theorem for locally compact groups on which the operator-valued lifting is based.","marker":"[58]"}],"fun_headline_variants":["Hao-Ng isomorphism proved for all locally compact groups","Reduced crossed products and Cuntz-Pimsner algebras commute","Hao-Ng theorem extended to every locally compact group","All locally compact groups: Hao-Ng holds for reduced products"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is the invariance principle invoked in Proposition 4.2 — that twisting a representation with the unique extension property by any group element again gives a representation with that property — which the paper cites to an unpublished note and another paper's proposition but does not prove under its non-unital hypotheses.","fun_headline_variants_meta":{"raw":{"variants":["Hao-Ng isomorphism proved for all locally compact groups","Reduced crossed products and Cuntz-Pimsner algebras commute","Hao-Ng theorem extended to every locally compact group","All locally compact groups: Hao-Ng holds for reduced products"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00072,"raw_usage":{"total_tokens":3200,"prompt_tokens":879,"completion_tokens":2321,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":495,"completion_tokens_details":{"reasoning_tokens":2251}},"tokens_in":495,"tokens_out":2321,"duration_ms":17247,"temperature":1.0,"reasoning_tokens":2251,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:38:48.005091+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a C*-dynamical system $(B,G,\\alpha)$, an $\\alpha$-invariant operator subalgebra $A$ generating $B$, and a representation $\\pi$ with the unique extension property with respect to $A$ such that $\\pi \\circ \\alpha_g$ fails that property for some $g \\in G$; then the uniqueness step of Proposition 4.2 collapses. The decisive test is whether $C^*_e(A \\rtimes_{\\alpha,r} G) \\cong C^*_e(A) \\rtimes_{\\alpha,r} G$ still holds in such an example, since a failure there would disprove the paper's central claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the invariance principle for the unique extension property that Proposition 4.2 invokes but does not prove."},{"cited_title":"2, 416–438","cited_arxiv_id":null,"evidence_quote":"Defines the unique extension property for non-unital operator algebras and contains the invariance discussion cited in the twisting step."},{"cited_title":"3, 257–270","cited_arxiv_id":null,"evidence_quote":"Provides the minimal S-projection argument used to realize the injective envelope inside a given injective range."},{"cited_title":"New series 8 (1982), no","cited_arxiv_id":null,"evidence_quote":"Constructs the Fubini tensor products that make the operator-valued Maharam lifting theorem work."},{"cited_title":"1, 226–233","cited_arxiv_id":null,"evidence_quote":"Identifies the Cuntz-Pimsner algebra of a correspondence with the C*-envelope of its tensor algebra."},{"cited_title":"258, American Mathematical Society, 2019","cited_arxiv_id":null,"evidence_quote":"Supplies the operator-algebra crossed product framework and the independence of the reduced crossed product from the chosen embedding."},{"cited_title":"2, 1160–1197","cited_arxiv_id":null,"evidence_quote":"Provides the hyperrigid version of the C*-envelope commutation and the proof strategy that this paper extends beyond hyperrigidity."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the equivariant lifting theorem for locally compact groups on which the operator-valued lifting is based."}],"review_version":1}