{"id":"cefda27e-561c-4f51-bdde-8d4fd70d13e0","arxiv_id":"2411.16051","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"For the FK-Ising model with q=2, the partition function describing interfaces conditioned on a one-arm event solves the radial BPZ equations and yields a rigorous radial Loewner driving function.","lead":"This paper proves that the scaling limit of FK-Ising random-cluster interfaces conditioned on a one-arm event is governed by a radial Schramm-Loewner evolution with drift given by a solution to the radial BPZ equations. It also builds a general family of positive radial BPZ solutions from global multiple SLE curves.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 2.8 is the load-bearing gap: the W-weighted partition functions are asserted to satisfy radial BPZ equations and a martingale property with a one-line proof that does not handle the non-local event W; Theorem 1.2's driving function depends on this.","rationale":"The reader's verdict is CONDITIONAL, with the weakest assumption identified as the uniform convergence in Corollary 3.8 and equation (3.13). That concern is real: the passage from discrete conditional one-arm probabilities to the continuum CR^{-1/8} weighting requires a joint convergence argument that is only sketched. I do not dispute it. However, I find a more fundamental gap earlier in the logical chain: Lemma 2.8, which supplies the continuum martingale property for the W-weighted partition functions used to define G^(1/8). The proof of Lemma 2.8 is omitted entirely, and the cited Lemma 2.2 does not address the indicator 1_W. The event W is non-local and depends on the connectivity of the remaining curves to the slit endpoints; its conditional expectation given the first curve's past is not obviously expressible through Z^(r)_{α;w} evaluated at the evolved marked points. A cascade relation is needed. This gap is load-bearing because Theorem 1.2's conclusion that the limiting interface is radial SLE with driving function (1.6) follows only after combining the discrete-to-continuum limit (3.10) with Lemma 2.8. The paper contains independent support in the detailed proof of Proposition 1.4 and the external results cited, so I do not regard the gap as a known counterexample, but it must be filled or annotated before the central claim is secure. The proposed cascade/N=1 check would settle whether the assertion in Lemma 2.8 is correct or whether the proof requires a genuinely new idea.","tokens_in":28238,"tokens_out":12664,"duration_ms":114862,"concrete_test":"Derive the cascade relation for Z^(r)_{α;w}: express Z^(r)_{α;w}(θ_1,...,θ_2N) as an expectation over the first chordal SLE of a product of Z^(r) on the two components of the slit domain, analogous to Lemma 2.6. If this relation holds, Lemma 2.8 follows by the same argument as Lemma 2.2. As a concrete check, set N=1 and compute Z^(r)_{α;w}(θ_1,θ_2) = (2 sin((θ_2−θ_1)/2))^{−2h} E[1_{z right of η} CR(U\\η;z)^{−r}] for a single chordal SLE_κ, either analytically by solving the resulting PDE or numerically by simulating the SLE, and verify the radial BPZ equation (1.12) with ℵ=(6−κ)(κ−2)/(8κ)−r. A failure for any r<1−κ/8 would refute Lemma 2.8 and hence Theorem 1.2; if the N=1 check passes, the missing cascade for general N must still be supplied.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Lemma 2.8 (Section 2.5) asserts that the W-weighted partition functions Z^(r)_{α;w} satisfy the radial BPZ equations and that, under Q^(r)_{α;w}, the first curve is radial SLE_κ weighted by Mt(Z^(r)_{α;w}). The proof is a single sentence: 'This can be proved in the same way as Lemma 2.2.' But Lemma 2.2's proof uses identity (2.18), which expresses the conditional expectation of CR^{-r} given the first curve's history as a ratio of full partition functions evaluated at the conformally evolved marked points. That identity relies on the conformal Markov property of global multiple SLE and the factorization of Z^(r)_α under the Loewner flow. The event W(η;z) in Definition 2.7 is a global topological condition on the entire curve configuration: it requires z to stay on a specified side of every curve that touches the component Ω_η(z). Conditional on the first curve's past, the remaining curves live in a slit domain whose boundary includes the two sides of the slit; whether W holds depends on how those remaining curves connect to the slit endpoints, not only on the conformal images of the original 2N marked points. Thus the asserted equality Mt(Z^(r)_{α;w}) = conditional expectation of 1_W CR^{-r} requires a cascade relation for Z^(r)_{α;w} analogous to Lemma 2.6, which is neither stated nor proved. Without Lemma 2.8, the mixture in (3.10) cannot be identified as radial SLE weighted by Mt(G^(1/8)), and the driving function (1.6) in Theorem 1.2 is unsupported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the radial BPZ equations associated with Dubédat's commutation relation and their connection to the scaling limit of FK-Ising interfaces conditioned on a one-arm event. It constructs positive solutions Z^(r)_α to the radial BPZ system (1.12) as global multiple SLE expectations of CR(Ω∖η;z)^{-r}, and introduces a W-weighted variant Z^(r)_{α;w} defined through the event that the interior point stays on prescribed sides of all relevant curves. The main result, Theorem 1.2, states that for the FK-Ising model q=2 (κ=16/3), the law of the collection of interfaces conditional on the one-arm event converges to global N-SLE weighted by CR^{-1/8}, equivalently that the first interface is a radial SLE_κ driven by (1.6), where the partition function G^(1/8) satisfies the radial BPZ equations (1.7). The proof combines known convergence results for FK-Ising interfaces [BPW21, FPW24], the one-arm asymptotic Lemma 1.5, and the BPZ construction of Section 2.","tokens_in":28621,"tokens_out":7180,"duration_ms":64742,"significance":"If correct, Theorem 1.2 is a substantial result: it gives the first rigorous scaling limit for multiple FK-Ising interfaces under a conditioning that effectively changes the partition function, and it identifies the limiting curve as radial SLE weighted by a positive solution of the radial BPZ system. The overall strategy is natural and the algebraic parts of the paper, such as the Itô calculus derivation in Lemma 2.2 and the hypoellipticity argument, are standard and appear sound. The paper also credits and cleanly uses the relevant prior inputs, including [BPW21, FPW24, KS16], and it provides an independent CLE-based proof of the known one-arm estimate in Lemma 1.5. However, as detailed in the major comments, two load-bearing points are only sketched: the proof of Lemma 2.8 and the joint convergence limit (3.13). These gaps prevent the manuscript from being fully convincing in its present form, although the central claim seems plausible and the gaps appear fixable.","major_comments":[{"comment":"Lemma 2.8 is asserted with the one-line proof 'This can be proved in the same way as Lemma 2.2,' but the W-weighted case is not a routine variant. Lemma 2.2 uses identity (2.18), which expresses the conditional expectation of CR(U∖η)^{-r} given η(1)[0,t] as a ratio of full partition functions evaluated at the conformally evolved marked points. For Z^(r)_{α;w}, the event W(η;z) from Definition 2.7 is a global topological condition involving the side of every curve touching the component Ω_η(z); conditional on the past of η(1), whether W holds depends on how the remaining curves connect to the two sides of the slit, not only on the images of the original marked points. An analogue of the cascade relation in Lemma 2.6 for the weighted partition functions is therefore needed, and it is neither stated nor proved. Since Definition 3.9 and equation (3.10) build G^(1/8) from Z^(r)_{α;w}, the driving function (1.6) in Theorem 1.2 is unsupported unless Lemma 2.8 is given a full proof.","section":"Section 2.5, Lemma 2.8"},{"comment":"The proof of Theorem 1.2 asserts the joint limit (3.13) with the instruction that it can be handled as in Corollary 3.8, but this is not a formality. Corollary 3.8's argument uses the epsilon-annulus truncation in (3.7)–(3.8) and dominated convergence for the indicator W(ηδ;zδ); passing to the version with an arbitrary bounded continuous test function F requires a uniform, curve-dependent version of Lemma 1.5 for the random domains Ω^δ_{ηδ(zδ)}, and the paper does not state or prove such a uniformity statement. If this uniformity fails, the identified limit in (3.10), and hence the driving function in Theorem 1.2, could be incorrect. This point should be replaced with a precise lemma and proof.","section":"Section 3.3, Eq. (3.13)"}],"minor_comments":[{"comment":"The word 'even' in 'We denote by W(η;z) the even that...' should be 'event'.","section":"Section 2.5, Definition 2.7"},{"comment":"The word 'crtitical' in 'We now consider the crtitical FK-Ising model' is a typo and should be 'critical'.","section":"Section 3.2, opening paragraph"},{"comment":"The sentence 'Comparing (A.8) with (A.10) gives (A.11)' appears to refer to (A.9) rather than (A.8); the displayed line (A.11) follows from (A.9) and (A.10).","section":"Appendix A, end of proof of Lemma 1.5"},{"comment":"The notation Q^ for the chordal SLE law used in Corollary 2.4 is not defined in the surrounding text; please introduce it explicitly.","section":"Section 2.1, Corollary 2.4"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and the high-level strategy is promising. The two gaps identified in the major comments are proof obligations rather than mere exposition choices, and they concern the central claim of Theorem 1.2. In particular, Lemma 2.8 and the joint convergence limit (3.13) need full proofs before publication. I see no grounds to reject the central claim itself; the issues appear fixable within the manuscript's scope."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the quick take: the paper proves a real new theorem—FK-Ising interfaces conditioned on the one-arm event converge to radial SLE with drift given by G^{(1/8)}—and, in the process, constructs positive radial BPZ solutions for arbitrary N by conformal-radius weighting of global multiple SLEs. The N=1 case was known [WW24]; the general construction and the identification with the FK-Ising one-arm exponent 1/8 are new. The main computational statements (Ito calculus, local martingale condition, hypoellipticity) look correct, and the exponent matches: for κ=16/3, r_1=1/8 gives ℵ = (16−κ^2)/(32κ) = −7/96.\n\nThe soft spots are all about completeness rather than correctness. Lemma 2.8 is the one that needs attention: it asserts that the W-weighted partition functions satisfy BPZ and give the weighted SLE martingale, with a one-line proof appealing to Lemma 2.2. The stress-test note worries that W is a non-local event and the conditional expectation isn't obviously a ratio of partition functions. I think the worry is answerable: W is conformally invariant, and after conditioning on the first curve's past, the image configuration is again a global N-SLE with the first point replaced by the tip, so E[1_W CR^{−r} | η[0,t]] = e^{rt} Z^w(...)/Z(...) follows exactly as in (2.18). But the paper should write this out; the one-liner is not enough for a result that carries the conclusion. Similarly, the limits (3.13) and the sketch in Corollary 3.8 rely on an epsilon-annulus + RSW argument that is standard but not detailed; (3.13) is asserted with 'one can proceed as in Corollary 3.8.' That's fillable but needs to be in the paper. Lemma 1.5 rederives a known result with a somewhat heavy appendix; this is mostly fine.\n\nWho gets value: people working on SLE classification, radial BPZ, and critical FK-Ising. The paper is a worthwhile contribution and deserves a serious referee. My recommendation: send it to peer review, but insist that the authors expand the proofs of Lemma 2.8 and the uniform limit (3.13) before acceptance. If those sketches hide a real obstruction, the theorem falls, but I don't see one.","headline":"A genuinely new conditional scaling limit for FK-Ising interfaces driven by radial BPZ solutions, with a sound main argument but a few proof sketches that must be expanded before the paper is complete.","tokens_in":29131,"tokens_out":9263,"would_cite":true,"duration_ms":80928,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["60J67","82B43"],"pacs":[],"model":"deepseek-v4-flash","headline":"For the FK-Ising model, an interface conditioned on a one-arm connection converges to a radial Loewner chain whose partition function solves the radial BPZ equations.","keywords":["radial BPZ equations","commutation relation","FK-Ising model","random-cluster model","one-arm event","radial Loewner chain","global multiple SLE","partition functions"],"falsifier":"Simulate the critical FK-Ising model in a unit disc with one interior point and two marked boundary points, condition on the one-arm event, extract the radial driving function of the interface, and compare its drift to $\\kappa\\,\\partial_1\\log G^{(1/8)}$ with $G^{(1/8)}$ given by (1.8); a statistically significant mismatch would show the conjectured BPZ solution is not the lattice limit.","tokens_in":28046,"feed_emoji":"🌀","tokens_out":15665,"duration_ms":121998,"temperature":0.7,"pith_summary":"This paper proves that, in the scaling limit of the critical FK-Ising model (cluster weight $q=2$, parameter $\\kappa=16/3$), an interface conditioned on a one-arm event—the event that a fixed interior point is connected by an open cluster to a specified boundary arc—is described by a radial Loewner chain (a growing family of conformal maps encoding a random curve) with an explicit Brownian driving function plus a drift built from a solution of the radial BPZ equations (a system of PDEs from conformal field theory). This confirms, for the FK-Ising case, a general conjecture about random interfaces in polygons with an interior point. The authors construct positive solutions to the radial BPZ system by taking global multiple SLEs (Schramm-Loewner evolutions) and weighting them with a power of the conformal radius of the component containing the interior point, then identify the lattice model's normalized conditional probabilities with these solutions. If the theorem is correct, it supplies a direct bridge from a lattice conditioning event to a conformal field theory equation, and a template that extends to other critical models once the relevant interface and loop convergence are known.","feed_headline":"One-arm FK-Ising interfaces follow a radial BPZ law","feed_subtitle":"Conditional interface limit in polygons has explicit driving function; the q=2 conjecture is now a theorem.","key_machinery":"The central object is the weighted global multiple SLE partition function $Z_\\alpha^{(r)}(\\Omega;x_1,\\ldots,x_{2N};z)=Z_\\alpha(\\Omega;x_1,\\ldots,x_{2N})E_\\alpha[\\operatorname{CR}(\\Omega\\setminus\\eta;z)^{-r}]$, together with the variant $Z_{\\alpha;w}^{(r)}$ that restricts to the event $W(\\eta;z)$ fixing which side of each interface contains the interior point. The paper shows that, for $\\kappa\\in(0,6]$ and $r<1-\\kappa/8$, these functions are positive solutions of the radial BPZ system (1.2) with $\\aleph=(6-\\kappa)(\\kappa-2)/(8\\kappa)-r$; the proof runs through the local martingale $M_t(Z)=g'_t(0)^{r-\\tilde h}\\prod_{j=2}^{2N}\\phi'_t(\\theta_j)^h Z(\\xi_t,\\phi_t(\\theta_2),\\ldots,\\phi_t(\\theta_{2N}))$ with $h=(6-\\kappa)/(2\\kappa)$ and $\\tilde h=(6-\\kappa)(\\kappa-2)/(8\\kappa)$. Zeroing the drift of this martingale gives the BPZ equation, and hypoellipticity upgrades weak solutions to smooth ones. On the lattice side, the same expression emerges from the domain Markov property of the FK-Ising model: conditioning on the one-arm event multiplies the interface law by the conditional connection probability, which converges to $\\operatorname{CR}(\\Omega\\setminus\\eta;z)^{-1/8}$.","core_discovery":"The central claim is Theorem 1.2: for critical FK-Ising interfaces in a polygon with alternating boundary conditions, the law of each interface $\\eta_j^\\delta$ conditional on the one-arm event $A^\\delta$ converges weakly, as the mesh $\\delta\\to0$, to the image under the conformal map $\\varphi^{-1}$ of the radial Loewner chain whose driving function solves (1.6), up to the first time a neighboring marked point is swallowed. The partition function $G^{(1/8)}$ in the drift is a positive solution of the radial BPZ equations (1.7) with constant $\\aleph=(16-\\kappa^2)/(32\\kappa)$, and for $N=1$ it has the explicit form $(\\sin((\\theta_2-\\theta_1)/2))^{1-6/\\kappa}(\\sin((\\theta_2-\\theta_1)/4))^{8/\\kappa-1}$. The proof identifies $G^{(1/8)}$ as the scaling limit of the normalized one-arm probabilities: the unconditioned interface law converges to a mixture of global multiple $\\mathrm{SLE}_{16/3}$ weighted by the appropriate meander connection probabilities, and conditioning multiplies each SLE by the indicator of the winding event $W$ times $\\operatorname{CR}(\\Omega\\setminus\\eta;z)^{-1/8}$, the conformal-radius power from the one-arm exponent. That weighted object is exactly the partition function for which Proposition 1.4 establishes radial BPZ equations.","pith_inferences":["The construction suggests the radial BPZ constant $\\aleph$ should be read as a tunable arm-exponent parameter: other conditioning events, such as two-arm or multi-arm connections, would correspond to other values of $r$ and hence other constants $\\aleph$ in the same family of solutions.","One could test the $N=1$ formula numerically: simulate the FK-Ising interface conditioned on the one-arm event, extract the driving function of the radial Loewner chain, and compare its drift with $\\kappa\\,\\partial_1\\log G^{(1/8)}$ from (1.8).","The joint-convergence step around (3.13) suggests that a sharper quantitative version of Lemma 1.5, with errors uniform over interface configurations, would be the natural next ingredient, both to complete the argument rigorously and to extend it to models without full CLE convergence.","The same weighted-global-SLE machinery may transfer to other planar critical random-cluster models ($q\\neq2$) as soon as the required interface and loop convergence results appear, making one-arm conditioning a general route from lattice probabilities to BPZ equations."],"forward_implications":["For $q=2$, $\\kappa=16/3$, the conditional interface has a precise weak limit described by the radial Loewner SDE (1.6), so scaling-limit statements about the conditioned FK-Ising interface can be read off from the SLE with this driving function.","The partition function $G^{(1/8)}$ solves radial BPZ with the specific constant $\\aleph=(16-\\kappa^2)/(32\\kappa)$, and for $N=1$ it is explicit, giving a closed-form prediction for the one-arm conditioned interface in a two-point domain.","The same strategy proves Conjecture 1.1 for Bernoulli site percolation on the triangular lattice ($\\kappa=6$), because convergence to SLE$_6$ and CLE$_6$ is already known.","For general $q\\in[1,4)$, Conjecture 1.1 would follow from the same argument once convergence of a single interface to SLE$_\\kappa$ and of the loop ensemble to CLE$_\\kappa$ are established."],"supporting_citations":[{"why":"Provides the convergence of unconditioned FK-Ising interfaces to global multiple SLE_{16/3}, used as Proposition 3.4.","marker":"[BPW21]"},{"why":"Supplies the limiting connection probabilities of multiple FK-Ising interfaces through the meander matrix, used as Proposition 3.6.","marker":"[FPW24]"},{"why":"Derives radial BPZ equations for N=1 radial SLE, the base case the paper extends to global N-SLE in Proposition 1.4.","marker":"[WW24]"},{"why":"Gives the one-arm exponent r1=(3κ−8)(8−κ)/(32κ) for CLE, fixing the power of the conformal radius in the partition function.","marker":"[SSW09]"},{"why":"Establishes the full scaling limit of FK-Ising interfaces as branching SLE, a key input for the CLE convergence used in Lemma 1.5.","marker":"[KS16]"},{"why":"Provides the convergence of boundary-touching FK-Ising loops to CLE_{16/3}, used in the proof of Lemma 1.5.","marker":"[KS19]"},{"why":"Constructs pure partition functions of global multiple SLE, the starting objects for the weighted partition functions in Proposition 1.4.","marker":"[Wu20]"}],"fun_headline_variants":["FK-Ising one-arm interfaces: radial BPZ driving law","Explicit scaling limit for one-arm FK-Ising interfaces","Radial BPZ yields exact one-arm FK-Ising interface law","One-arm FK-Ising limit is BPZ-driven Loewner chain"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument rests on the assertion, sketched around (3.13), that the normalized conditional one-arm probabilities converge uniformly to $\\operatorname{CR}(\\Omega\\setminus\\eta;z)^{-1/8}$ jointly with the interface curves; if this joint convergence fails, the identified driving function may not be the correct limit.","fun_headline_variants_meta":{"raw":{"variants":["FK-Ising one-arm interfaces: radial BPZ driving law","Explicit scaling limit for one-arm FK-Ising interfaces","Radial BPZ yields exact one-arm FK-Ising interface law","One-arm FK-Ising limit is BPZ-driven Loewner chain"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000246,"raw_usage":{"total_tokens":1516,"prompt_tokens":895,"completion_tokens":621,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":544}},"tokens_in":511,"tokens_out":621,"duration_ms":5731,"temperature":1.0,"reasoning_tokens":544,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:36:40.777395+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Simulate the critical FK-Ising model in a unit disc with one interior point and two marked boundary points, condition on the one-arm event, extract the radial driving function of the interface, and compare its drift to $\\kappa\\,\\partial_1\\log G^{(1/8)}$ with $G^{(1/8)}$ given by (1.8); a statistically significant mismatch would show the conjectured BPZ solution is not the lattice limit.","supporting_citations":[],"review_version":1}