{"id":"c0144755-a248-4c26-a2e0-913423ff850b","arxiv_id":"2507.06283","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Conformal bootstrap bounds for U(1)-charged scalars in 3d are saturated by the CP^2 model's large-N and lattice predictions, suggesting the CP^2 deconfined quantum critical point is a conformal field theory.","lead":"This paper uses the conformal bootstrap, a mathematical technique for constraining quantum field theories, to study the CP^2 model, a leading candidate for the simplest deconfined quantum critical point in three dimensions. The authors find that the bootstrap's allowed region is saturated by the model's predicted operator dimensions, yielding new estimates for the spectrum of monopole operators.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The CP^2 identification depends on inputting Δ1=0.755 from large-N extrapolation, but the paper never varies Δ1 over the range allowed by conflicting lattice estimates, so the claimed saturation could be an artifact of that choice.","rationale":"The reader's verdict is CONDITIONAL, and I agree with that verdict. My single most load-bearing concern is the undetermined sensitivity of the bootstrap output to the externally input Δ1, which the reader also flagged as one of two fragile premises; the reader additionally emphasized the assumed relevant-operator spectrum. I focus on Δ1 because a shift within the existing lattice range directly changes all extracted dimensions and could destroy the claimed matching, whereas a relevant SU(3) adjoint would affect the completeness of the characterization but would not necessarily change the U(1)-sector data. The paper is honest about limitations, including Appendix C, and the numerical methods are standard and carefully documented, but no code is shipped and the EFM estimates are non-rigorous. The proposed concrete test is computational and uses the same setup and data files already referenced in the paper. If the test shows strong sensitivity to Δ1, the central claim would need to be weakened; if not, the concern does not land. Because the reader's conditional verdict already accommodates this uncertainty, no change to the verdict is required.","tokens_in":18537,"tokens_out":11863,"duration_ms":138064,"concrete_test":"Rerun the Navigator minimization of Δ0 at fixed Δ1 for Δ1 = 0.71, 0.74, 0.755, 0.77, and 0.785, using the same spectral assumptions, correlators (φ0, φ1, φ2), spin sets, and Λ = 19, 21, ..., 31 as in Tables IV-VI, then apply the same linear 1/Λ extrapolation used for Table I. Compare the extrapolated Δ0, Δ2, Δ3, and Δ4 to the large-N values (1.81, 3.10, 4.59) and to the lattice Δ0 estimates (1.46(7), 1.28). If the outputs stay within the quoted errors across the full Δ1 range, the Δ1 input is not load-bearing. If Δ0 moves by more than about 0.1, or Δ2, Δ3, Δ4 deviate from large-N by more than their quoted errors as Δ1 goes from 0.71 to 0.785, then the CP^2 identification is an artifact of the chosen input and the verdict should be weakened.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing unsecured step is the input Δ1=0.755. This number comes from the large-N expansion of CP^{N-1} extrapolated to N=3, but the paper assigns it no error and does not test the bootstrap output against the full range of lattice determinations: Lou et al. [14] quote Δ1=0.71(4), Harada et al. [15] quote 0.785. Since Δ1 is a free input to the Navigator minimization, moving it changes the boundary point and hence all extracted Δq. The paper shows only the slice at Δ1=0.755. The q=2,3,4 'matches' in Table I are comparisons to the same large-N framework that produced Δ1; they are internal consistency checks, not independent anchors. The only independent CP^2 anchor is Δ0 against lattice [14], and that comparison is marginal (1.61(1) vs 1.46(7), roughly 2σ) and conflicts with [15] (1.28). If the true Δ1 were 0.785, the minimized boundary could move the extracted Δ0 toward 1.3 and shift Δ2, Δ3, Δ4 away from large-N values, undermining the claim that the bootstrap bound describes CP^2. The paper's own Appendix C shows the same procedure for CP^4 produces a boundary point whose Δ4 (3.5(4)) is far from large-N (6.21), so identifying boundary points with the target theory is not automatic. Because the sensitivity of the extracted CFT data to Δ1 is unquantified, the central claim is conditional on one specific value of an extrapolated input.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper applies the conformal bootstrap to 3d CFTs with O(2) global symmetry, using correlators of scalar operators with U(1) charges q=0,1,2, to study the CP^{N-1} model at N=3 (CP^2). After assuming that only the lowest q=0,1,2 scalar operators are relevant and setting Δ1=0.755 from a large-N extrapolation, the authors minimize Δ0 and read off the spectrum at the resulting boundary point. They report Δ2=1.841(1), Δ3=3.173(4), Δ4=4.65(9), and Δ0=1.61(1), claiming agreement with large-N predictions for monopole operators and with a lattice estimate for Δ0 from Ref. [14]. They also extract spinning monopole dimensions Δ_{q,ℓ} for ℓ≤4 and compare them to a large-charge effective theory for ℓ≤q. The paper concludes that the critical CP^2 model is described by this bootstrap bound.","tokens_in":18874,"tokens_out":4924,"duration_ms":51926,"significance":"If the identification is correct, this is an important step: it would provide bootstrap evidence that CP^2 is a CFT with a single relevant U(1)-singlet scalar and would yield predictions for a hierarchy of monopole operators. The numerical work is carefully documented: parameter tables (Tables IV and V), finite-Λ data (Tables VI, VIII, X), explicit extrapolation procedures (Appendices B and C), and an auxiliary data file. The comparison to the O(2) large-charge effective theory (Appendix D) is a useful cross-check. The paper is honest about the non-rigorous nature of extremal functional extractions and the weak SU(3) sector. However, the central claim rests on load-bearing assumptions—the value of Δ1, the irrelevance of q=3,4 scalars, and the interpretation of boundary saturation—that are not yet fully tested. The result is best viewed as a suggestive identification rather than a conclusive one.","major_comments":[{"comment":"The central boundary point is computed at the single value Δ1=0.755, taken from the large-N extrapolation of Ref. [17] extrapolated to N=3. The paper never varies Δ1 over the range suggested by conflicting lattice determinations (Δ1=0.71(4) in Ref. [14] and Δ1=0.785 in Ref. [15]). Since Δ1 is a free input in the Navigator minimization, all extracted quantities (Δ0, Δ2, Δ3, Δ4, and spinning dimensions) can depend on it. A sensitivity scan is essential: if the boundary point moves significantly with Δ1, the claimed agreement shown in Table I could be an artifact of the chosen input. The only independent CP^2 anchor, Δ0, matches lattice Ref. [14] only at about the 2σ level (1.61(1) vs 1.46(7)) and is far from the Ref. [15] value (1.28). The authors should provide a scan over the allowed Δ1 range or otherwise quantify how robust their outputs are to this input.","section":"Section III, Figure 2, Table I"},{"comment":"The same procedure, applied to CP^4 with Δ1=1.005, yields Δ4(bootstrap)=3.5(4), which is in stark disagreement with the large-N value 6.21, while Δ3=4.6(4) vs 4.18 is only marginal. This demonstrates that a boundary point at the large-N Δ1 does not by itself identify the target theory. The authors do not address why the CP^4 failure is not expected to affect the CP^2 result. Since the CP^2 identification rests on the same logic of inputting Δ1 and minimizing Δ0, this inconsistency must be resolved—for example, by explaining why the relevant operator content is different for CP^2 (so that the imposed gap structure is justified) or by showing that the CP^4 discrepancy is due to an unsupported spectral assumption.","section":"Appendix C, Table IX"},{"comment":"The bootstrap input explicitly imposes that all q=3 and q=4 scalar operators are irrelevant, i.e., that the target theory has exactly one relevant operator per charge q=0,1,2. This assumption is not derived from the U(1) sector and is not tested by the correlators considered; a relevant q=3 scalar (as in the critical O(2) model, which the paper excludes) or a relevant SU(3) adjoint would alter the allowed region and the location of the boundary point. The U(1)-only bootstrap cannot rule out these alternatives. Because the central identification depends on this assumption, it should be treated as a hypothesis and either supported by additional mixed correlators or explicitly framed as such in the conclusions.","section":"Section III, paragraph beginning 'We next consider correlators'"}],"minor_comments":[{"comment":"The caption lists '∆1 = 0.1005'; this should be '∆1 = 1.005'.","section":"Table X caption"},{"comment":"The statement that increasing the gap above Δ0 above three 'might address this problem' for larger N is vague and speculative; the authors should either provide a concrete estimate for the needed gap or remove the conjecture.","section":"Section IV, paragraph 'We would also like to generalize'"},{"comment":"The text describes the Δ0 comparison with lattice Ref. [14] as a 'match', but the numbers 1.61(1) vs 1.46(7) differ by about two combined standard deviations; the discrepancy should be stated explicitly.","section":"Table I and Section III"},{"comment":"The abstract and Section I promise results for N=4,5, but only CP^4 (N=4) bootstrap output is shown in Appendix C; CP^5 appears only in Figure 1. Please clarify whether CP^5 bootstrap results were obtained.","section":"Appendix C and Introduction"},{"comment":"The column header 'spin-ranges' should be 'spin sets' or 'spin ranges' for grammatical clarity.","section":"Table IV"},{"comment":"The phrase 'the N=3 bosonic theory' refers to CP^2 (i.e., CP^{N-1} with N=3); defining this notation earlier would help the reader avoid confusion with the N in CP^{N-1}.","section":"Introduction, first paragraph"}],"recommendation":"major_revision","confidential_remarks":"The paper is an interesting and carefully executed numerical bootstrap study. The main concern is that the central claim is conditional on untested assumptions: the sensitivity to Δ1, the spectal assumption about relevant operators, and the contradictory CP^4 result. A sensitivity analysis and a deeper discussion of the CP^4 discrepancy would be needed before publication. The authors should also temper the conclusion from 'suggests' to a more hedged statement given the marginal lattice agreement."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe short version: this is a well-executed bootstrap paper that gives the first numerical bootstrap treatment of the N=3 CP model, and it is honest about what it cannot prove. The central claim—that the critical CP^2 model sits on the boundary of the allowed U(1) CFT region—is conditional on taking Δ1=0.755 from the large-N expansion and on assuming only the lowest q=0,1,2 scalars are relevant. The authors say that in so many words, so the paper is not overselling.\n\nWhat is genuinely new: they run the O(2) correlator system with q=0,1,2, get access to q=3,4, and produce spinning monopole dimensions (Table II). The match between bootstrap and large-charge effective theory for ℓ≤q is a nice, non-trivial pattern, and the same pattern in O(2) (Appendix D) strengthens it. The numerical documentation is careful—parameters in Tables IV/V, Λ-dependence in Tables VI/VIII/X, leave-p-out error bars, and the auxiliary data file QED3 data.nb is provided. That is real work.\n\nThe soft spots are real but proportionate. The stress-test point is the main one: Δ1 is a free input, set to 0.755 without error, and the lattice spread (0.71(4) to 0.785) is not explored. Since the boundary point and all extracted dimensions move with Δ1, the 'matches' in Table I are partly internal consistency checks with the same large-N framework that supplied Δ1. The only independent anchor, Δ0 vs lattice [14], is marginal and conflicts with [15]. The CP^4 appendix is a useful caution: the same procedure gives Δ4 ≈ 3.5(4) versus large-N 6.21, so boundary saturation does not automatically identify the theory. And the relevant-operator assumption is imposed, not derived; a relevant q=3 operator would break the identification. These are not fatal, but they keep the paper at the level of strong evidence, not proof.\n\nI don't buy the circularity objection: Δ1 comes from prior large-N work, and the outputs are compared to, not fitted to, lattice and large-N. That is a legitimate cross-check.\n\nWho should read it: DQCP and bootstrap people will want it. It deserves a serious referee. Recommendation: send to peer review.","headline":"A careful bootstrap study that plausibly identifies the CP^2 critical point, but the central claim rests on an unvaried large-N input and an assumed relevant spectrum.","tokens_in":19448,"tokens_out":2817,"would_cite":true,"duration_ms":29924,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that the CP$^2$ model, the simplest remaining candidate for a deconfined quantum critical point, is described by a conformal bootstrap bound, with scaling dimensions matching large-$N$ and lattice results.","keywords":["conformal bootstrap","deconfined quantum critical point","CP^2 model","monopole operators","large-N expansion","large charge expansion","scaling dimensions","O(2) global symmetry"],"falsifier":"A lattice or Hamiltonian simulation of the would-be CP$^2$ critical theory that found a relevant scalar operator of U(1) charge $3$, a relevant SU(3) adjoint, or a value of $\\Delta_1$ clearly outside $0.755$ would break the matching. Concretely, measuring the lowest charge-3 excitation and the singlet dimension $\\Delta_0$ with errors small enough to distinguish $\\Delta_0\\approx 1.61$ from the competing lattice estimate $\\Delta_0\\approx 1.28$ would settle whether the bootstrap point is the physical theory.","tokens_in":18288,"feed_emoji":"⚛️","tokens_out":9353,"duration_ms":84342,"temperature":0.7,"pith_summary":"The paper aims to show that the CP$^2$ model, the $N=3$ member of the CP$^{N-1}$ family, is the simplest remaining candidate for a deconfined quantum critical point, meaning a continuous transition between ordered phases described by an emergent gauge field and a conformal field theory. By studying four-point functions of scalar operators with U(1) charges $0$, $1$, and $2$, and assuming these are the only relevant operators, the authors find that fixing the charge-$1$ dimension to its large-$N$ value and minimizing the charge-$0$ dimension produces scaling dimensions for charges $2$, $3$, and $4$ that match large-$N$ monopole calculations, and a charge-$0$ dimension that matches one lattice estimate. The paper also predicts the dimensions of the lowest spinning monopole operators and finds they agree with the large-charge effective theory for spin less than or equal to charge. If the identification is right, the bootstrap supplies a concrete operator spectrum for the simplest deconfined quantum critical point, including testable lattice predictions.","feed_headline":"Bootstrap pins the CP^2 critical point to a bound","feed_subtitle":"Crossing symmetry yields operator dimensions matching large-N monopole and lattice predictions.","key_machinery":"The central object is the conformal bootstrap for mixed correlators of scalar operators $\\phi_0$, $\\phi_1$, $\\phi_2$ with charges $0$, $1$, $2$ under the U(1), or O(2), global symmetry. Crossing symmetry of the four-point functions, combined with unitarity and the assumed spectrum, produces a space of allowed scaling dimensions; at the boundary of this space one finds an approximate solution to crossing from which operator data can be extracted. The $q=0,1,2$ external operators give access to exchanged operators of charges up to $4$. The input $\\Delta_1$ comes from a large-$N$ saddle-point computation of monopole operator dimensions via the state-operator correspondence, and the comparison for spinning operators uses the large-charge effective theory formula $\\Delta_{q,\\ell}=c_{3/2} q^{3/2}+c_{1/2} q^{1/2}-0.0937+\\sqrt{\\ell(\\ell+1)/2}+O(q^{-1/2})$ with coefficients fixed by large $N$. The numerical bootstrap machinery includes a truncation parameter $\\Lambda$ whose extrapolation to infinity controls the reported errors.","core_discovery":"On the paper's own terms, the central discovery is that the critical CP$^2$ model appears on the boundary of the allowed region of three-dimensional conformal field theories with O(2) global symmetry and a single relevant operator of each charge $q=0,1,2$. Setting $\\Delta_1=0.755$, the value from the large-$N$ expansion extrapolated to $N=3$, and minimizing $\\Delta_0$, the bootstrap yields $\\Delta_2=1.841(1)$, $\\Delta_3=3.173(4)$, $\\Delta_4=4.65(9)$, and $\\Delta_0=1.61(1)$. These numbers agree with the large-$N$ monopole dimensions $\\Delta_2=1.81$, $\\Delta_3=3.10$, $\\Delta_4=4.59$ and with the lattice value $\\Delta_0=1.46(7)$ from [14], while a competing lattice estimate [15] gives $\\Delta_0=1.28$ and $\\Delta_1=0.785$. The lowest spinning monopole dimensions computed from the bootstrap match the large-charge effective theory for $\\ell\\leq q$, the same pattern seen in the critical O(2) model. The paper concludes that this suggests the critical CP$^2$ model is described by the bootstrap bound.","pith_inferences":["The paper's success in the U(1) sector does not by itself certify the full SU(3) structure; the authors note the adjoint bootstrap gives weak bounds, so a mixed-correlator study involving SU(3) adjoints is the natural next test.","If the $\\ell\\leq q$ matching holds generally, then large-charge effective theory may be a reliable spectral tool even at small charge and moderate spin; computing non-lowest monopoles at large $N$ would test this directly.","The bootstrap's spectral assumption could be probed by lattice searches for a charge-3 relevant operator; absence of such an operator would support the CP$^2$ identification, while presence would point to the critical O(2) model.","A natural extension is to apply the same U(1)-sector bootstrap to gauge theories with Chern-Simons couplings or QCD3, though the paper does not carry this out."],"forward_implications":["If correct, CP$^2$ is a conformal field theory with a single relevant U(1)-singlet scalar, and Tables I and II give its lowest scalar and spinning monopole scaling dimensions.","The bootstrap prediction $\\Delta_0\\approx 1.61$ provides a target for lattice simulations that can distinguish it from the earlier estimate $\\Delta_0\\approx 1.28$.","The match for spinning monopoles at $\\ell\\leq q$ suggests the large-charge effective theory works beyond its formal regime $\\ell\\ll q^{1/2}$, as also seen in the critical O(2) model.","The same U(1)-sector bootstrap, with $\\Delta_1$ replaced by the large-$N$ value for larger $N$, gives partial results for CP$^3$ and CP$^4$, but the CP$^{N-1}$ model no longer sits on the lower bound as $N$ grows.","A relevant $q=3$ operator would instead indicate the critical O(2) model, so the assumed spectrum is what selects CP$^2$ from other O(2) conformal field theories."],"supporting_citations":[{"why":"Supplies the large-$N$ monopole scaling dimensions for $q=1,2,3,4$ that provide the input $\\Delta_1=0.755$ and the comparison values.","marker":"[17]"},{"why":"Gives the lattice estimate $\\Delta_0=1.46(7)$ that the bootstrap $\\Delta_0$ is matched against.","marker":"[14]"},{"why":"Provides the O(2) bootstrap setup and the crossing equations for correlators of charge $0,1,2$ operators.","marker":"[23]"},{"why":"Gives the earlier allowed-region bootstrap for O($N$) in the $(\\Delta_1,\\Delta_0)$ plane, generalized here to larger $\\Delta_1$.","marker":"[39]"},{"why":"Fixes the large-charge effective theory coefficients using large $N$, used for the spinning monopole predictions.","marker":"[26]"},{"why":"Derives the large-charge expansion formula for operator dimensions at large global charge.","marker":"[24, 25]"},{"why":"Establishes the state-operator correspondence for monopole operators and the saddle-point calculation of their dimensions.","marker":"[16]"},{"why":"Provides the numerical optimization method used to scan the allowed region and minimize $\\Delta_0$.","marker":"[37]"}],"fun_headline_variants":["Bootstrap nails CP^2 to a single bound","CP^2 critical point sits on bootstrap bound","Bootstrap boundary is the CP^2 critical point","CP^2 fits bootstrap bound for monopole and lattice"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The identification of the bound with the CP$^2$ model rests on the assumptions that the theory has exactly one relevant scalar of each charge $q=0,1,2$ and none with higher charge, and that the large-$N$ extrapolated value $\\Delta_1=0.755$ used as input is accurate.","fun_headline_variants_meta":{"raw":{"variants":["Bootstrap nails CP^2 to a single bound","CP^2 critical point sits on bootstrap bound","Bootstrap boundary is the CP^2 critical point","CP^2 fits bootstrap bound for monopole and lattice"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000853,"raw_usage":{"total_tokens":3764,"prompt_tokens":1057,"completion_tokens":2707,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":673,"completion_tokens_details":{"reasoning_tokens":2644}},"tokens_in":673,"tokens_out":2707,"duration_ms":21755,"temperature":1.0,"reasoning_tokens":2644,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T19:08:21.739353+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A lattice or Hamiltonian simulation of the would-be CP$^2$ critical theory that found a relevant scalar operator of U(1) charge $3$, a relevant SU(3) adjoint, or a value of $\\Delta_1$ clearly outside $0.755$ would break the matching. Concretely, measuring the lowest charge-3 excitation and the singlet dimension $\\Delta_0$ with errors small enough to distinguish $\\Delta_0\\approx 1.61$ from the competing lattice estimate $\\Delta_0\\approx 1.28$ would settle whether the bootstrap point is the physical theory.","supporting_citations":[{"cited_title":"Antiferromagnetic to valence-bond-soild transitions in two-dimensional SU(N) Heisenberg models with multi-spin interactions","cited_arxiv_id":"0908.0740","evidence_quote":"Gives the lattice estimate $\\Delta_0=1.46(7)$ that the bootstrap $\\Delta_0$ is matched against."},{"cited_title":"The large charge expansion at large N","cited_arxiv_id":"1805.00501","evidence_quote":"Fixes the large-charge effective theory coefficients using large $N$, used for the spinning monopole predictions."},{"cited_title":"Monopoles in CP(N-1) model via the state-operator correspondence","cited_arxiv_id":"0809.2816","evidence_quote":"Establishes the state-operator correspondence for monopole operators and the saddle-point calculation of their dimensions."}],"review_version":1}