{"id":"896bbaf4-7f6a-4f6e-aaaf-a83128ca6b45","arxiv_id":"2411.09354","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Nonlinear scattering functionals on Lorentzian manifolds with an asymptotically Minkowskian infinity determine the manifold's topology, differentiable structure, conformal type, and the nonlinear coefficient up to a conformal transformation.","lead":"The paper proves that measurements of how small waves scatter off a nonlinear wave equation on a curved space-time can reveal the space-time's topology and geometry up to a conformal change. This gives a rare case where nonlinear scattering data solves an inverse problem that is still open for linear waves.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Equation (99) in the proof of Theorem 3 identifies a nonlinear scattering functional with the future trace of a linear source solution; this equality is generally false, though the proof appears repairable by using the nonlinear equation (40) in place of (93).","rationale":"The reader's weakest_assumption identifies exactly the same step: Eq. (99) of Theorem 3. I agree that this is the most load-bearing point, because Theorem 3 is the hinge that converts scattering data into the near-field source-to-solution maps that all later microlocal reconstruction steps use. My reading sharpens the concern slightly: the equality is not just unproven; it is false as written, because the scattering functional is defined through the nonlinear Goursat problem and the function u in (93) solves a linear equation. For small incoming fields of amplitude ε, the nonlinear correction is of order ε^κ, so the asserted equality fails at leading nontrivial order unless the nonlinear coefficient vanishes on the wave's support. That would invalidate the proof as written. However, I also see a straightforward repair: start directly with the nonlinear equation (40) instead of the linear equation (93). Since the nonlinear coefficient A and the potential B vanish on the non-physical past N^-, the solution in N^- is the same standard linear wave for both manifolds; hence the I^- traces agree. The assumed equality of scattering functionals then gives agreement of the future traces on I^+, and uniqueness for the linear Goursat problem in N^+ gives equality of the full source-to-solution maps. Thus the main theorem is plausible and the submitted proof needs a correction, not a rejection of the whole approach. The reader's CONDITIONAL verdict is therefore the right one, and my read does not change it.","tokens_in":61859,"tokens_out":10486,"duration_ms":111589,"concrete_test":"Concrete check: in Minkowski space, take κ=4 and a smooth nonzero A supported in the physical region N. Let u=εu0 be the solution of the linear equation (93) with a small source in N^-, with u0 positive on an open set, and set h^-=u|I^-. Compute the scattering functional to order ε^4 via S(h^-)(q)=εu0(q)-ε^4∫G(q,y)A(y)u0(y)^4dy+O(ε^7), where G is the retarded propagator. For any q for which the integral is nonzero (which is generic), Eq. (99) is false; a single exact or numerical evaluation with a Gaussian u0 settles whether the equality can hold.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The weakest point is Eq. (99) in Section 3.2, proof of Theorem 3. There, u^(j) is defined by the linear problem (93), (□_{g_ext}+B)u^(j)=f, and its trace h^-=u^(j)|I^- is then inserted into the scattering functional. The proof asserts u^(1)|I^+(t2)(q') = S^(1)_{t1,q}(u^(1)|I^-)(q') = S^(2)_{t1,q}(u^(2)|I^-)(q') = u^(2)|I^+(t2)(q'). But S_{t1,q}(h^-) is defined in Definition 4 as the value at q of the solution of the nonlinear Goursat problem (21)-(23). The nonlinear solution with past radiation h^- is not the linear solution u^(j): for an incoming field of size ε, the difference is of order ε^κ. In a Minkowski model with □u+A u^κ=0, S(h^-)(q)=u(q)-∫ G(q,y)A(y)u(y)^κ dy+O(ε^{2κ-1}), which is not identically u(q) when A is nonzero. Thus Eq. (99) is not merely under-justified; it is generally false. Since Theorem 3 is the only bridge from scattering data to the near-field source-to-solution maps used in the rest of the paper, this is a load-bearing gap. A repair is available: replace (93) by the nonlinear equation (40). Because A and B vanish on N^-, the solution in N^- is the same standard linear wave for both manifolds, so the I^- traces coincide; equality of scattering functionals then gives equality of future traces, and the Goursat problem in N^+ yields equality of the source-to-solution maps. The submitted proof therefore needs correction rather than being fundamentally unsound.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper introduces a new type of scattering data, called non-linear scattering functionals, for semilinear wave equations on globally hyperbolic Lorentzian manifolds with an asymptotically Minkowskian infinity. It claims (Theorem 2) that these functionals uniquely determine the topology, differentiable structure, and conformal type of the manifold, as well as the Lorentzian metric and the coefficient of the nonlinearity up to the multiplicative transformations in (34). The proof strategy is to reduce the scattering problem to a near-field problem on an extended manifold obtained by gluing non-physical regions N^+ and N^- to the Penrose compactification, and then to apply the higher-order linearization method from previous work [75, 83]. Theorem 3 is the bridge that converts equality of scattering functionals into equality of source-to-solution maps, and the rest of the paper develops the microlocal analysis of nonlinear interactions to reconstruct the conformal geometry and the coefficient a(x). The paper also contains examples (wormholes, locally Schwarzschild horizons, FLRW-type space-times) and a discussion of generalizations to non-smooth metrics, with details deferred to later work.","tokens_in":62230,"tokens_out":5930,"duration_ms":61108,"significance":"If the proof is repaired, this would be a substantial advance in inverse scattering for nonlinear hyperbolic equations on Lorentzian manifolds. The introduction of scattering functionals as data that remain well-defined even when classical scattering operators fail due to blow-up is a useful conceptual contribution, and the reduction from scattering data to near-field source-to-solution maps via Penrose compactification is an elegant strategy that extends the earlier local methods of [75, 83] to global inverse scattering. The paper also gives concrete and nontrivial examples of manifolds with several ends or event-horizon-like structures that fall within the scope of the theorem. The authors are explicit about the external benchmark results they rely on, and they acknowledge the complementary work of Hintz, Sá Barreto, Uhlmann, and Zhang [53].","major_comments":[{"comment":"The equality S^{(j)}_{t1,q}(u^{(j)}|_{I^-})(q') = u^{(j)}|_{I^+}(t2)(q') is asserted without proof and is generally false. The scattering functional S_{t1,q} is defined in Definition 4 as the value at q of the solution of the nonlinear Goursat problem (21)-(23), which contains the term A e{u}^\\kappa. The function u^{(j)} appearing in (93) solves the linear equation (□_{gext}+B)u^{(j)} = f, not the nonlinear equation. For an incoming trace h^- of small amplitude ε, the nonlinear solution with this past radiation field differs from the linear solution by a term of order ε^\\kappa when A is nonzero. Thus the identification in (99) is not a consequence of the definitions; it would only hold in the trivial case A ≡ 0. Since this equality is the only step that transfers the assumed equality of scattering functionals to equality of the future traces of the waves used to construct the near-field source-to-solution maps, Theorem 3 is not proved as written.","section":"Section 3.2, Eqs. (93)-(100)"},{"comment":"A repair of the gap in Eq. (99) is available but requires a nontrivial modification of the proof. One should replace the linear problem (93) by the nonlinear problem (40), i.e., solve (□_{gext}+B)w + A w^\\kappa = f with supp(w) ⊂ J^+(supp(f)). Because A and B vanish on N^-, the solution w in N^- is the same standard linear wave for both manifolds, so the traces w^{(1)}|_{I^-} and w^{(2)}|_{I^-} coincide and lie in B^-(R(t1)) with sufficiently small norm by Theorem 5(ii). The assumed equality of scattering functionals then gives equality of w^{(1)}|_{I^+} and w^{(2)}|_{I^+}, and the Goursat problem in N^+ (now with the same nonlinear equation) yields equality of the source-to-solution maps L_{g^j,B_j,A_j,p^+,K_n}. The submitted proof does not make this replacement, and the current line of argument is therefore not merely under-justified but incorrect at the displayed equality.","section":"Section 3.2, Eqs. (93)-(100)"}],"minor_comments":[{"comment":"In the sentence defining the compact sets K_n, the text reads 'compact sets that care closures of open sets'; the word 'care' should be 'are'.","section":"Section 1.5"},{"comment":"The symbol ε is used both for the radius of the domain D^{(ε)}(S_{t1,q}) in (33) and as the smallness parameter in Theorem 5; the relation between the two (for instance, the choice of ε(t1,q1) in the sentence after (33)) would be clearer if denoted by different letters.","section":"Definition 4"},{"comment":"In the proof that Σ∪{i0} is a Cauchy surface of Next, the assertion that J^-_{Next}(x1) ∩ J^+_{Next}(i0) is compact is used without proof; this is not immediate from the definition of Next and should be justified or replaced by an explicit argument.","section":"Section 2.0.2, proof of Lemma 1"},{"comment":"The subsection 'On the further generalizations' contains a sketch for metrics that are only C^0 near i0 and states that details will be presented elsewhere; this is acceptable as a remark, but it should be clearly labeled as work in progress rather than as part of the main theorem, since the main theorem is restricted to the smooth case.","section":"Section 5.0.1"},{"comment":"The statement of Theorem 2 and the transformation (34) do not mention the linear potential d(x) that appears in the equation (18). The authors might add a remark explaining whether d is also recovered, or whether it is an obstruction to uniqueness that the statement intentionally omits.","section":"Theorem 2"}],"recommendation":"major_revision","confidential_remarks":"The paper is a serious and interesting contribution, but the proof of Theorem 3 has a load-bearing gap at Eq. (99). The repair outlined in the report (replacing the linear equation (93) by the nonlinear equation (40)) appears feasible and would preserve the overall strategy, so I do not recommend rejection. The authors should be asked to revise the proof accordingly and to clarify the handling of the smallness conditions for the traces in the repaired argument."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper has a genuinely new idea and a strong main theorem, but the submitted proof has a load-bearing step in Theorem 3 that is currently false as written. I think it is repairable, and the paper deserves serious refereeing, but not in its present form.\n\nWhat is new: the scattering functionals are defined even when the classical scattering operator is not, because they evaluate the future radiation field at a point q before any possible blow-up. The reduction of far-field scattering data to near-field source-to-solution maps via the extended Penrose compactification is also new, and the class of manifolds—open globally hyperbolic with asymptotically Minkowskian ends, non-trivial topology, several infinities—is much broader than what linear inverse scattering can handle. Theorem 2 gives a clean uniqueness statement up to the expected conformal transformations. The paper is honest about relying on [75] and [83] for the near-field inverse problem; that is acceptable, though it makes the verification burden heavy.\n\nThe soft spot is in Section 3.2, proof of Theorem 3. Equation (99) asserts that S_{t1,q}(u^(j)|I^-) equals u^(j)|I^+ for u^(j) solving the linear equation (93). That is not a consequence of the definitions: S_{t1,q} is defined through the nonlinear Goursat problem (21)-(23), and for an incoming trace of size ε the nonlinear solution differs from the linear one by terms of order ε^κ. So the equality is generally false, not just missing justification. Since Theorem 3 is the only bridge between scattering functionals and the near-field source-to-solution maps used in the rest of the paper, this is load-bearing.\n\nThe good news is that the fix seems clear: replace (93) by the nonlinear equation (40). Because A and B vanish on N^-, the I^- traces of these nonlinear solutions coincide with the linear ones, so (96) still holds. Equality of scattering functionals then gives equality of the future traces, and the Goursat problem in N^+ yields the equality of source-to-solution maps. So the proof should be correctable, but as submitted it needs that correction.\n\nWho this is for: anyone working on inverse problems for nonlinear hyperbolic equations, scattering theory on Lorentzian manifolds, or using nonlinear interactions as a tool. The scattering functional data and the Penrose reduction will be useful beyond this paper. I would send it to peer review—this is exactly the kind of paper referees should engage with—but with a clear request to fix Theorem 3.","headline":"Genuinely new inverse scattering result with a repairable but load-bearing gap in Theorem 3 (Eq. (99)).","tokens_in":62756,"tokens_out":2771,"would_cite":true,"duration_ms":26913,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["35R30","35L05","35L71","53C50"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that nonlinear scattering functionals uniquely determine the topology, differentiable structure, and conformal type of a globally hyperbolic Lorentzian manifold with an asymptotically Minkowskian infinity, and recover the…","keywords":["inverse scattering","semilinear wave equation","Lorentzian manifolds","scattering functionals","conformal type","nonlinear interaction of waves","Penrose compactification","asymptotically Minkowskian infinity"],"falsifier":"In Minkowski space with a small, positive, compactly supported coefficient $a$, take an incoming radiation field that is the past trace of a linear wave generated by a compact source in the non-physical past, and compare, to first order in the amplitude, the future trace of the nonlinear Goursat solution with the linear future trace. Their difference is essentially $-\\int G(q,y) A(y) u_{\\mathrm{lin}}(y)^\\kappa\\, dy$; if a suitable source makes this integral nonzero, the asserted equality $S_{t_1,q}(u|_{I^-})=u|_{I^+}$ fails and the reduction in Theorem 3 breaks.","tokens_in":61653,"feed_emoji":"📡","tokens_out":14491,"duration_ms":146242,"temperature":0.7,"pith_summary":"This paper proves an inverse scattering uniqueness theorem for the semilinear wave equation $\\Box_g u + d u + a u^\\kappa = 0$ with $\\kappa \\geq 4$ on a globally hyperbolic Lorentzian manifold that has at least one asymptotically Minkowskian infinity visible from the whole spacetime. The data are nonlinear scattering functionals: for each small incoming radiation field supported on a compact piece of past null infinity and each observation point $q$ on future null infinity, one records the value of the outgoing radiation field at $q$. The paper shows that these functionals determine the topology, differentiable structure, and conformal type of the spacetime, and that the metric and the nonlinear coefficient $a$ are recovered up to the rescaling $g \\mapsto e^{2\\gamma}g$, $a \\mapsto e^{(\\kappa-3)\\gamma}a$. This matters because the scattering functionals remain well defined even when incoming waves blow up before a classical scattering operator could be defined, and the corresponding inverse problem for the linear wave equation remains open.","feed_headline":"Small nonlinear waves determine spacetime up to scale","feed_subtitle":"New scattering functionals recover topology, conformal class, and the nonlinear coefficient even when waves blow up.","key_machinery":"The load-bearing object is the scattering functional $S_{t_1,q}(h^-)=h^+(q)$, defined for incoming radiation fields $h^-$ in the class of concentrated waves as the value at a point $q$ of the future radiation field of the nonlinear Goursat problem with data on past null infinity. The argument's engine is the reduction of these far-field point evaluations to near-field source-to-solution maps on the extended Penrose spacetime $N_{\\mathrm{ext}} = N \\cup N^+ \\cup N^-$; a central step is the analysis of $\\kappa$-th-order interactions of four conormal waves, whose interaction produces a wave whose leading singularity is supported on the future light cone of the interaction point and whose principal symbol contains $A(q)$ as a factor. That structure lets the proof reconstruct earliest light observation sets and then the conformal type of the metric.","core_discovery":"The central claim is Theorem 2: two Lorentzian manifolds with asymptotically Minkowskian infinities whose scattering functionals agree for all small incoming radiation fields are related by a diffeomorphism $\\Psi$ and a conformal factor $e^{2\\gamma}$, and their nonlinear coefficients satisfy $a^{(1)} = e^{(\\kappa-3)\\gamma}\\Psi^* a^{(2)}$. In other words, far-field nonlinear scattering data uniquely determine the manifold's topology, its differentiable structure, the conformal class of its metric, and the nonlinear coefficient up to the gauge freedom in (34). The proof conformally compactifies the physical spacetime, glues non-physical past and future regions onto the compactification, and first shows that the scattering functionals determine source-to-solution maps on this extended spacetime; nonlinear interactions of waves inside the physical region then act as structured probes whose scattered light cones reveal the conformal geometry.","pith_inferences":["Beyond the paper: because the theorem recovers only the conformal class, any practical reconstruction must fix a gauge, for example by normalizing the metric at one point or prescribing the conformal factor; the theorem itself does not address that choice.","Beyond the paper: the scattering functionals are point evaluations of the outgoing field, so the result suggests that a dense family of point receivers at future null infinity carries the same information as the full radiation field, which would ease the demands on scattering experiments.","Beyond the paper: the closing sketch indicates the same strategy may extend to positive-ADM-mass, Schwarzschild-like spacetimes and cubic nonlinearities; if carried out, inverse scattering would reach more physical black-hole exteriors, where the direct problem is already harder."],"forward_implications":["Equal scattering functionals force a conformal diffeomorphism between the two spacetimes, so the causal and null-cone structure of the manifold is fully encoded in the far-field data.","Once a conformal gauge is chosen, the coefficient $a$ of the nonlinear term is uniquely determined, because the remaining freedom is exactly the multiplication by $e^{(\\kappa-3)\\gamma}$ accompanying the metric rescaling.","Scattering functionals provide meaningful inverse-scattering data in regimes where the classical scattering operator is not defined, including cases where some incoming waves lead to finite-time blow-up.","The reduction to near-field source-to-solution maps on the extended spacetime means that local-measurement reconstruction techniques can be applied to purely far-field scattering measurements.","For conformally FLRW spacetimes satisfying the no-particle-horizon condition, restricted scattering data already determine the metric and nonlinear coefficient in the reconstruction domain, a setting the paper connects to cosmological models."],"supporting_citations":[{"why":"Method the paper extends: higher-order linearization and reconstruction of conformal type from earliest light observation sets, invoked as Theorem 1.2 and Theorem 4.5.","marker":"[75]"},{"why":"Supplies the microlocal analysis of kappa-th-order nonlinear interactions and the recovery of the nonlinear coefficient A from source-to-solution maps.","marker":"[83]"},{"why":"Provides the inverse-problem framework for nonlinear hyperbolic equations with disjoint sources and receivers that the extended-spacetime reduction adapts.","marker":"[39]"},{"why":"Gives the uniqueness theorem for the characteristic Cauchy (Goursat) problem used to ensure scattering functionals are well defined.","marker":"[94]"},{"why":"Establishes strong Huygens' principle on the 3-sphere, used to define the concentrated-wave radiation-field classes B plus and B minus and their support properties.","marker":"[84]"},{"why":"Defines the Penrose conformal compactification that turns the spacetime infinities into boundaries and underlies the extended manifold N_ext.","marker":"[99]"},{"why":"Provides the characteristic Cauchy problem well-posedness used in Theorem 3 to propagate future radiation data back into the non-physical future region.","marker":"[59]"}],"fun_headline_variants":["Nonlinear scattering data fix spacetime up to scale","Blow-up waves still recover manifold topology and scaling","Inverse scattering on curved spacetimes: unique up to scaling","Even exploding waves pin down conformal geometry","Nonlinear wave interactions map spacetime structure"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reduction to near-field measurements assumes that a scattering functional, which is defined through the nonlinear equation, acts like the identity on the incoming trace of a linear wave: it returns the linear future trace unchanged, even though the nonlinear term is present along the wave's path.","fun_headline_variants_meta":{"raw":{"variants":["Nonlinear scattering data fix spacetime up to scale","Blow-up waves still recover manifold topology and scaling","Inverse scattering on curved spacetimes: unique up to scaling","Even exploding waves pin down conformal geometry","Nonlinear wave interactions map spacetime structure"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000653,"raw_usage":{"total_tokens":2981,"prompt_tokens":921,"completion_tokens":2060,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":537,"completion_tokens_details":{"reasoning_tokens":1987}},"tokens_in":537,"tokens_out":2060,"duration_ms":15100,"temperature":1.0,"reasoning_tokens":1987,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T20:45:04.454676+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"In Minkowski space with a small, positive, compactly supported coefficient $a$, take an incoming radiation field that is the past trace of a linear wave generated by a compact source in the non-physical past, and compare, to first order in the amplitude, the future trace of the nonlinear Goursat solution with the linear future trace. Their difference is essentially $-\\int G(q,y) A(y) u_{\\mathrm{lin}}(y)^\\kappa\\, dy$; if a suitable source makes this integral nonzero, the asserted equality $S_{t_1,q}(u|_{I^-})=u|_{I^+}$ fails and the reduction in Theorem 3 breaks.","supporting_citations":[{"cited_title":"Kurylev, M","cited_arxiv_id":null,"evidence_quote":"Method the paper extends: higher-order linearization and reconstruction of conformal type from earliest light observation sets, invoked as Theorem 1.2 and Theorem 4.5."},{"cited_title":"Lassas, G","cited_arxiv_id":null,"evidence_quote":"Supplies the microlocal analysis of kappa-th-order nonlinear interactions and the recovery of the nonlinear coefficient A from source-to-solution maps."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the uniqueness theorem for the characteristic Cauchy (Goursat) problem used to ensure scattering functionals are well defined."},{"cited_title":"Lax.Hyperbolic Partial Differential Equations","cited_arxiv_id":null,"evidence_quote":"Establishes strong Huygens' principle on the 3-sphere, used to define the concentrated-wave radiation-field classes B plus and B minus and their support properties."},{"cited_title":"Oksanen, M","cited_arxiv_id":null,"evidence_quote":"Defines the Penrose conformal compactification that turns the spacetime infinities into boundaries and underlies the extended manifold N_ext."},{"cited_title":"Hörmander","cited_arxiv_id":null,"evidence_quote":"Provides the characteristic Cauchy problem well-posedness used in Theorem 3 to propagate future radiation data back into the non-physical future region."}],"review_version":1}