{"id":"853435a9-80f6-4b3c-893f-4e019ad8ace0","arxiv_id":"2607.20395","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Comparison, maximal-flow existence, and generic uniqueness for Hele-Shaw-type free boundary problems with sign-changing velocity are proved in general dimension.","lead":"This analysis paper proves a comparison principle, existence of maximal flows, and generic uniqueness for a moving-boundary (Hele-Shaw) problem in any dimension, even when the boundary can shrink. The framework is a new notion of 'viscosity flows' and matters for PDE models of tumors, crowds, and porous media.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 1.1's proof requires a positive-separation hypothesis not stated; with touching initial sets (3.13) fails, so the comparison principle is unproven as written.","rationale":"The paper's central claim is Theorem 1.1, and the proof's opening estimate (3.13) implicitly requires a positive-distance condition not stated in the theorem. This is an internal gap: even if Assumptions 1.3 and 1.5 hold (including the reader's flagged Assumption 1.5(i)), the comparison proof fails for initial sets that touch. I therefore focus on this rather than on Assumption 1.5(i), which is a conditional assumption the authors explicitly flag as open for general F. Both are real; the proof gap is more directly load-bearing because it affects the theorem's validity under its own stated hypotheses. The concrete example with intervals shows (3.13) is false when dist(Ω1(0),Ω2(0)^c)=0. The conclusion of Theorem 1.1 may still be true in this example, so the paper is likely fixable by adding compact containment or an approximation argument; hence the reader's CONDITIONAL verdict remains appropriate. Agreement is partial: the reader mentioned this issue as point (1) in the rationale but selected Assumption 1.5(i) as the weakest assumption; I think the proof gap at (3.13) is the more load-bearing concern.","tokens_in":25601,"tokens_out":10176,"duration_ms":82431,"concrete_test":"Take d=1, F=−u″−1 (so p(x;U)=−(x−a)(x−b)/2 for U=(a,b)), V≡c>0, and define Ω1(t)=[−ct,1+ct], Ω2(t)=[−ct,2+ct]. These are exact constant-speed viscosity flows with Ω1(0)=[0,1]⊂[0,2]=Ω2(0) but dist(Ω1(0),(Ω2(0))^c)=0. Evaluate (3.13) at t=0: it requires (−r,1+r)⊆(r,2−r) for r=ε²h>0, which fails for every r>0. This directly falsifies the assertion that (3.13) follows from the hypotheses. If the authors did not intend compact containment, the proof must be modified; if they did, Theorem 1.1 should state Ω1(0)⋐Ω2(0).","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing flaw is in the proof of the comparison principle, not in an unverified assumption. At the start of the proof of Theorem 1.1, after setting r=ε²h and δ=εh, the authors claim (3.13): B_r(Ω^{δ,r,h}_1)∩(R^d×[0,t])⊆Ω^{δ,r,h}_2 for all small t, saying this follows from Ω1(0)⊂Ω2(0) and small h. But at t=0, the left side is an r-neighborhood of Ω1(0) and the right side is the r-interior of Ω2(0). The hypothesis Ω1(0)⊂Ω2(0) does not imply dist(Ω1(0),(Ω2(0))^c)>0. For example, with d=1, Ω1(0)=(0,1) and Ω2(0)=(0,2), inclusion is strict but the distance is 0; the r-neighborhood (−r,1+r) is never contained in the r-interior (r,2−r) for any r>0. Consequently (3.13) cannot hold no matter how small h is. The proof as written establishes comparison only under an extra compact-containment hypothesis Ω1(0)⋐Ω2(0), which is absent from Theorem 1.1. Since Theorem 1.2 and Theorems 4.1–4.2 rely on Theorem 1.1, this gap affects the paper's main well-posedness claims. It is likely repairable either by adding the missing hypothesis or by a limiting/approximation argument, but neither appears in the manuscript.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a set-based viscosity theory for a Hele-Shaw-type free boundary problem in arbitrary space dimension: an elliptic equation inside the positive phase and a free boundary velocity V(x,∇p)|∇p|. The main objects are viscosity subflows and superflows (Definition 2.3), defined by testing with regular space-time sets. The paper's central result is a comparison principle for such flows (Theorem 1.1), proved via space-time set regularizations (Lemmas 3.3–3.6) and an interposition lemma from [7]. From comparison the authors obtain a Perron-style existence of maximal flows (Theorem 1.2) and two uniqueness results: one for initial sets whose boundary velocity has a definite sign (Theorem 4.1), and one showing generic uniqueness in the sense of a residual set of initial data (Theorem 4.2). The paper also connects the set-based definition to the more traditional function-based viscosity solution definition (Lemma 2.8), and it verifies the main structural assumptions for several linear elliptic operators.","tokens_in":25904,"tokens_out":11412,"duration_ms":97841,"significance":"If correct, this is a substantial contribution: it brings comparison and well-posedness for Hele-Shaw-type flows to general dimensions while allowing the free boundary velocity to be sign-changing, and it gives a workable bridge between set-testing and function-testing notions. The strategy is genuinely different from earlier work: set convolutions replace sup/inf-convolutions, and Lemma 3.5–3.6 encode the needed strict advancement. The paper is also honest about what is assumed: Assumption 1.5 is verified only for linear examples, and Lemma 2.8 is explicitly one-directional. No fitted parameters or numerical claims appear; the proofs are structured and mostly checkable. However, the proof of the comparison principle contains a concrete initial-separation gap, described below, that affects the paper's main well-posedness theorems as currently stated.","major_comments":[{"comment":"The proof of Theorem 1.1 requires an initial positive separation that is not present in the hypotheses. After setting r=ε²h and δ=εh, the text asserts that for h sufficiently small, B_r(Ω^{δ,r,h}_1)∩(R^d×[0,t])⊆Ω^{δ,r,h}_2 for all sufficiently small t. At t=0, the left-hand side is the r-neighborhood of (a time-rescaled) Ω1(0), while the right-hand side is the r-interior of Ω2(0). The hypothesis Ω1(0)⊂Ω2(0) does not imply dist(Ω1(0),∂Ω2(0))>0. For example, in d=1 with Ω1(0)=(0,1) and Ω2(0)=(0,2), inclusion is strict but (−r,1+r) is never contained in (r,2−r) for any r>0. Thus (3.13) fails no matter how small h is. The proof as written establishes comparison only under an extra compact-containment hypothesis. Since Theorem 1.2 and Theorems 4.1–4.2 rely on Theorem 1.1, this gap is load-bearing for the main well-posedness claims. It is likely repairable either by adding dist(Ω1(0),∂Ω2(0))>0","section":"§3.2, Eq. (3.13)"},{"comment":"The reduction to the non-degenerate case ∇ϕ≠0 at boundary contact points is not fully justified. The proof asserts that because Ω_ε(t) satisfies an interior ball condition, the Hopf maximum principle implies that the associated sub-function p_ε is non-degenerate near the boundary, and hence any test function touching p_ε from above at a boundary point must have nonzero spatial gradient. No Hopf lemma for viscosity solutions in merely interior-ball domains is stated or cited, and Ω_ε(t) is only a union of balls, so its boundary need not be C^{1,1}. This is a gap in one of the paper's advertised contributions (the flow/solution bridge). It does not directly affect the set-flow comparison, but the lemma should either be proved with a precise Hopf-type statement or the needed regularity/geometry assumption should be added.","section":"§2.2, Lemma 2.8, Step 2"}],"minor_comments":[{"comment":"The statement says 'the regular from the interior of Ω^δ_1' where it should be Ω^δ_2; this is presumably a typo.","section":"Lemma 3.3(ii)"},{"comment":"The notation uses (y1,s1) and (x1,t1) in the statement, but the intended points are (y2,s2) and (x2,t2); please correct.","section":"Lemma 3.6(ii)"},{"comment":"The state space C_r is defined as closed sets with an interior ball condition, while Theorem 1.2 is stated for open bounded sets. The relationship between the closed-set formulation and the open-set existence theorem should be clarified, e.g., by taking interiors or by adjusting the definitions.","section":"§4.3, Theorem 4.2"},{"comment":"The paper would be easier to use if it explicitly stated that the main theorems are conditional on Assumption 1.5(i)–(ii), which are verified only for linear elliptic operators and one-dimensional examples. This is not a defect, but the scope of the theory should be phrased carefully in the introduction and abstract.","section":"Assumption 1.5"}],"recommendation":"major_revision","confidential_remarks":"I recommend major revision rather than rejection: the main gap is concrete but appears repairable within the manuscript's framework. If the authors add the missing separation hypothesis to Theorem 1.1 or supply a limiting argument, and clarify the Hopf step in Lemma 2.8, the paper would be a strong contribution. The set-convolution approach and the generic uniqueness result are valuable, and the manuscript is otherwise carefully written."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Main takeaway: the comparison principle, which is the engine of the paper, is not proven for the hypotheses stated in Theorem 1.1. The proof of (3.13) needs dist(Ω1(0), (Ω2(0))^c)>0, and that is not in the assumptions. With touching initial sets, e.g. Ω1(0)=(0,1), Ω2(0)=(0,2) in d=1, (3.13) is false for every r>0. So as written the argument only gives comparison under compact containment Ω1(0)⋐Ω2(0). That is a load-bearing gap, not a cosmetic one: Theorem 1.2 and both uniqueness theorems lean on Theorem 1.1.\n\nThat said, the substance is real. The paper genuinely extends [38] from 1D to all dimensions, allows V to sign-change, and assembles: set-convolution regularization, interposition from Cardaliaguet-Ley, residual-set generic uniqueness, and a one-direction bridge between set-testing and function-testing viscosity notions. Kim's negative-velocity uniqueness needed Laplacian plus star-shaped or slope conditions; Cardaliaguet-Ley needed translation-invariant operators; this is the first combination covering general elliptic F in general dimensions with possibly negative velocity. The literature review in Remarks 1.7–1.9 is honest about the debts. No fitted parameters, no invented entities, and Assumption 1.5(ii) is explicitly flagged as open for general F, with only linear operators verified.\n\nSoft spots beyond (3.13): Assumption 1.5(i) is a strong regularity hypothesis—it postulates boundary regularity of ∂_t p and ∇p for every C^{1,1} set, and it is the entry point for Definition 2.3. It is verified for linear operators, but not derived from ellipticity; if it fails, the whole framework is undefined. The authors flag this, but it deserves as much scrutiny as (3.13). There are also minor typos: (3.1)'s definition of Ω^δ_2 appears to refer to Ω_1 rather than Ω_2, and Lemma 3.3(ii) says 'interior of Ω^δ_1' where it should say Ω^δ_2. The abstract's word 'well-posedness' is a bit strong given uniqueness is proved only for sign-definite initial velocity and for generic initial sets, but that is stated precisely in the theorems.\n\nVerdict: conditional, with a decent chance the comparison gap is repairable. But a referee needs to see the fix, not just be told it is straightforward. This deserves serious peer review; send it out.","headline":"Comparison theorem not proven as stated—needs positive separation at t=0—but the framework is novel and the gap looks repairable; deserves refereeing.","tokens_in":26484,"tokens_out":2742,"would_cite":true,"duration_ms":22327,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35B51","35R35","76D27"],"pacs":[],"model":"deepseek-v4-flash","headline":"A comparison principle governs Hele-Shaw type free boundaries in every spatial dimension, allowing the boundary to move inward or outward.","keywords":["Hele-Shaw flow","free boundary problem","viscosity solutions","viscosity flows","comparison principle","well-posedness","negative velocity","generic uniqueness"],"falsifier":"Take a C^{1,1} domain and an elliptic F satisfying Assumption 1.3, then compute the associated function's non-tangential boundary gradient; if one can find a pair where this gradient fails to be continuous, Assumption 1.5(i) is violated and the definition of viscosity flow has no value at boundary points, so the comparison principle cannot be stated. Constructing such a counterexample would settle that the framework collapses at its entry point.","tokens_in":25383,"feed_emoji":"🌊","tokens_out":4851,"duration_ms":37223,"temperature":0.7,"pith_summary":"This paper establishes that a broad class of Hele-Shaw type free boundary problems — where an elliptic equation holds inside a positive set and the boundary moves with a prescribed normal velocity — is well-posed in every dimension. The central result is a comparison principle: a viscosity flow that starts inside another must remain inside it. From that principle, the authors obtain existence of a maximal flow for any reasonable initial set, uniqueness when the initial boundary velocity is strictly inward or outward, and uniqueness for 'almost every' initial set. The theory is built on a set-based notion of viscosity solutions, which works even when the boundary velocity is negative, i.e. when the interface contracts.","feed_headline":"Hele-Shaw flow gets a comparison principle in any dimension","feed_subtitle":"Set-based viscosity theory proves uniqueness even when the interface shrinks.","key_machinery":"The central objects are viscosity flows: space-time sets of openness defined by testing against C^{1,1} sets from inside and outside. Each test set Φ has an associated function p(x; Φ(t)), the unique viscosity solution of the elliptic problem with zero boundary data; its gradient on the boundary supplies the normal velocity in the test inequality. The proof of comparison combines a set interpolation lemma that separates two disjoint sets by a C^{1,1} surface, with a set regularization based on space-time sup/inf convolutions that strictly advances or retreats the boundary, enabling a contradiction argument.","core_discovery":"The paper proves a comparison principle (Theorem 1.1) for viscosity flows of (1.1). Under structural assumptions on the elliptic operator F and the velocity V, if Ω1 is a viscosity subflow, Ω2 a viscosity superflow, and Ω1(0) ⊂ Ω2(0), then Ω1 ⊆ Ω2. This is used in a maximality construction to obtain a maximal viscosity flow (Theorem 1.2), and then to prove two uniqueness results: a unique flow when the initial normal velocity is strictly positive or strictly negative (Theorem 4.1), and a generic uniqueness result stating that for a residual set of initial domains the flow is unique (Theorem 4.2). The work also establishes a bridge between the set-testing notion of viscosity flow and the func","pith_inferences":["The comparison principle likely opens the way toward homogenization and stochastic homogenization results in higher dimensions, extending the authors' one-dimensional work to general settings.","The generic uniqueness result may be strengthened to full uniqueness if Assumption 1.5(i) can be verified for a wider class of nonlinear operators; currently the proof depends on that regularity hypothesis.","The set-based framework could adapt to other front propagation laws, including nonlocal or spatially dependent normal velocities, since the key ingredients are geometric rather than equation-specific."],"forward_implications":["Comparison gives uniqueness of viscosity flows starting from the same initial set when the initial free boundary velocity does not vanish (Theorem 4.1).","Existence of a maximal flow holds for any open bounded initial set satisfying the interior ball condition (Theorem 1.2), so the flow evolution is well defined for a broad class of initial data.","Generic uniqueness holds: for a residual set of initial domains in the Hausdorff metric, the flow is unique (Theorem 4.2).","The set-function bridge (Lemma 2.8) allows translating between two established notions of weak solution for Hele-Shaw type problems.","The theory covers negative free boundary velocities, corresponding to shrinking interfaces, which previous existence results for expanding flows did not address."],"fun_headline_variants":["Hele-Shaw comparison principle works for shrinking interfaces","Viscosity flows yield uniqueness of Hele-Shaw in all dimensions","Hele-Shaw well-posedness via comparison for arbitrary dimension","Shrinking Hele-Shaw interfaces: comparison proves uniqueness","Comparison principle handles negative velocity in Hele-Shaw"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"Assumption 1.5(i) — that for every space-time set with C^{1,1} boundary the associated elliptic solution has a continuous non-tangential gradient on the boundary — is the load-bearing premise: the definition of viscosity flow evaluates the velocity at that gradient, and the comparison proof needs this gradient to behave continuously under perturbations.","fun_headline_variants_meta":{"raw":{"variants":["Hele-Shaw comparison principle works for shrinking interfaces","Viscosity flows yield uniqueness of Hele-Shaw in all dimensions","Hele-Shaw well-posedness via comparison for arbitrary dimension","Shrinking Hele-Shaw interfaces: comparison proves uniqueness","Comparison principle handles negative velocity in Hele-Shaw"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000764,"raw_usage":{"total_tokens":3141,"prompt_tokens":571,"completion_tokens":2570,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":315,"completion_tokens_details":{"reasoning_tokens":2485}},"tokens_in":315,"tokens_out":2570,"duration_ms":15829,"temperature":1.0,"reasoning_tokens":2485,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T09:59:26.639672+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a C^{1,1} domain and an elliptic F satisfying Assumption 1.3, then compute the associated function's non-tangential boundary gradient; if one can find a pair where this gradient fails to be continuous, Assumption 1.5(i) is violated and the definition of viscosity flow has no value at boundary points, so the comparison principle cannot be stated. Constructing such a counterexample would settle that the framework collapses at its entry point.","supporting_citations":[],"review_version":1}