REVIEW 2 major objections 4 minor 33 references
On the inverse problem of the two-velocity tree-like graph
T0 review · 2 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash
Pith's one-line read The reduced Titchmarsh-Weyl matrix at all but one boundary vertex of a planar tree of strings determines the tree, its angles, and the two densities on every edge.
desk verdict Variable-density extension has a solid core but the main uniqueness theorem fails due to an exact gauge symmetry on the unmeasured root edge. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The machinery is the reduced Titchmarsh-Weyl matrix function, a matrix of $2\times 2$ blocks whose entry $M_{ij}(\lambda)$ maps the boundary value at leaf $j$ of the two-component displacement to its normal derivative at leaf $i$ for solutions of the spectral problem with Dirichlet conditions elsewhere. Around it the proof builds a leaf-peeling reduction: the star-graph step uses the matrix identities $A_k = S_{ki}A_i(S_{ki})^{-1}$ and $C_i = S_{in}B_n(S_{in})^{-1}$, where $S_\alpha$ is the rotation matrix between edge frames and $B_n$ collects the root-edge impedances $1/\sqrt{\varrho_n(0)}$, $1/\sqrt{\mu_n(0)}$, to pass from known boundary edges to the unknown interior edge and its incidence angle. The boundary control method supplies the locality needed to recover the density on each boundary edge independently from the diagonal entries $R_{ii}(t)$.
What would settle it
Compute the reduced Titchmarsh-Weyl matrix for a three-edge star graph, then repeat the computation after rotating the unobserved edge by 90 degrees in the plane and interchanging its two density functions; exact identity of the two matrices for all $\lambda$ would show that the angle and the density assignment on that edge are not individually determined by the stated data.
Extended reading notes
Core claim
The central claim, stated as Theorem 2, is that an arbitrary finite tree with a two-velocity wave equation on each edge is determined by the reduced Titchmarsh-Weyl matrix function $M_{ij}(\lambda)$ for $1\le i,j\le m-1$, i.e., by spectral data at every leaf except one chosen as the root. From those data one recovers the combinatorial tree, the edge lengths, the angles between neighbouring edges, and the two positive $C^2$ densities on every edge. The argument treats the root as clamped and works inward: first the boundary-control method extracts each boundary edge's densities from the diagonal of the response function, then the star-graph analysis recovers the transmission coefficients and the impedance of the next inner edge, and a leaf-peeling step eliminates the resolved leaves and produces the reduced data for the remaining tree. Iterating yields the full graph. The same conclusion is transferred to the dynamical setting in Theorem 3, where the reduced response operator on a finite time interval of length twice the controllability time determines the system.
Load-bearing premise
The proof assumes one can tell which of the two impedances at the root end of the excluded edge is the longitudinal one; otherwise a simultaneous 90-degree rotation of that edge and swap of its two densities leaves all the reduced data unchanged.
Editorial extensions
If this is right
- If Theorem 2 is right, an inspection of a tree from all but one boundary point is sufficient: the missing leaf can remain clamped and unobserved.
- Planar geometry becomes identifiable: the angles between edges at a vertex, not just the graph's combinatorial type, are encoded in the spectral data.
- The dynamical version gives a finite-time procedure: the reduced response operator on $[0,2\tau]$, where $\tau$ is the maximal optical distance from the excluded leaf to the others, determines the system.
- The recovery is constructive and layer-by-layer; Remark 2 asserts the method leads to a stable, numerically implementable algorithm, with numerical experiments already available for the one-velocity star case.
Reading between the lines
- The proof's Step 4 determines the excluded edge's incidence angle only up to a 90-degree rotation with a simultaneous swap of the two densities on that edge; unless an ordering or genericity assumption is imposed, the uniqueness statement in Theorem 2 is stronger than the argument supports.
- Because Theorem 3 is derived from Theorem 2 through the response-operator/TW correspondence, the same 90-degree-swap ambiguity would transfer to the finite-time dynamical inverse problem.
- The ambiguity disappears if the two impedances at the root end of the excluded edge are known to be unequal and their longitudinal/transverse assignment is fixed; alternatively, a single extra measurement at the excluded leaf would break the symmetry.
- The layer-by-layer scheme suggests a concrete numerical test: run the star-graph inversion on synthetic data with the excluded edge's endpoint densities nearly equal and watch the recovered angle and density assignment degrade as the two values approach each other.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper addresses the inverse problem of recovering a planar tree-like network of elastic strings with two wave speeds from the reduced Titchmarsh-Weyl matrix or the reduced dynamical response operator measured at all but one boundary vertex. The claimed results include unique recovery of the graph (edge lengths, topology, angles between edges) and of the two spatially varying densities on every edge. The proof proceeds by solving a star-graph inverse problem via the boundary control method and then applying a leaf-peeling procedure to arbitrary trees.
Significance. If the central uniqueness theorem were correct, the paper would represent a meaningful step beyond earlier constant-coefficient results and would provide a constructive algorithm with potential numerical implementation. The paper honestly engages with the boundary control method and builds on the authors' prior work. However, the main result as stated is false: the reduced data are invariant under a gauge transformation on the unmeasured root edge that swaps its two densities and rotates its local frame by 90 degrees. This is an exact counterexample, not a mere proof gap. A generic ordering assumption (or an explicit labeling convention) could salvage the result, but the theorems in the current form overclaim.
major comments (2)
- [Section 4, Step 4, Eq. (36)] The recovery of B_n and α_in from C_i = S_{in} B_n S_{in}^{-1} is ambiguous. Equation (36) determines C_i, whose trace and determinant provide only the unordered pair {1/sqrt(ρ_n(0)), 1/sqrt(μ_n(0))}. The reduced data do not identify which eigenvalue belongs to the longitudinal channel. Consequently, replacing (ρ_n, μ_n, α_in) by (μ_n, ρ_n, α_in+π/2) on the unmeasured edge leaves the global PDE system, the continuity conditions (3), and the force balance (4) pointwise invariant, because the rotation S_{π/2} conjugates the diagonal density matrix to the swapped one. Thus every M_{ij}(λ), 1≤i,j≤m−1, is identical for the two systems. If ρ_n(0)=μ_n(0), the angle α_in is completely unobservable. This invalidates Theorem 1 and Theorem 2 as stated. A generic ordering assumption (e.g., ρ_i(x)<μ_i(x) on every edge) or an explicit labeling of the longitudinal versus transverse channel is required.
- [Section 5, leaf-peeling step] The peeling argument propagates the ambiguity described above. When a sheaf is removed, the internal edge e_0 becomes the new boundary of the reduced tree, but its recovered density assignment and angle are known only up to the swap/π/2 transformation. The formulas for the reduced TW matrix fM(λ), in particular equations (44)-(48), use S_{10} and D_0, so the reconstructed data for the smaller tree inherit the same ambiguity. Therefore the iterative procedure cannot break the symmetry, and the uniqueness claim of Theorem 2 fails unless an additional ordering assumption is imposed on every edge.
minor comments (4)
- [Title and Abstract] The name Titchmarsh is misspelled as 'Titchmarch' throughout the manuscript; it should be 'Titchmarsh-Weyl'.
- [Section 3, Eq. (16)] The Fourier transform convention connecting M(k^2) and R(t) is not fully specified; for λ off the real axis the integral should be understood in a distributional sense, and a brief clarification would help the reader.
- [Section 5, Step 1] The definitions of pre-sheaf and sheaf are given informally; a precise topological definition (in terms of the subtree consisting of a vertex and its incident edges) would improve readability.
- [Theorem 3 proof] The proof of Theorem 3 is only a sketch, relying on controllability results from [29] and on an asserted equivalence between the reduced response and the reduced TW matrix. Since this theorem is a main advertised result, a more complete proof or a precise reference would be expected.
Circularity Check
No significant circularity: the recovery argument is constructive and relies on independent BC-method results; the main caveats are a missing ordering assumption in Step 4 and an explicitly sketched proof of Theorem 3, neither of which is a circular reduction.
full rationale
Walking the derivation chain, no step reduces to its own input. The reduced TW matrix M(λ) is the data, and the paper reconstructs densities, lengths, angles and connectivity by the BC method and leaf peeling. The one-interval recovery of ϱ_i, μ_i and the reflection/transmission coefficients (Eqs. (21)–(30), Step 3) is imported from the standard BC literature ([11,16,17] and related), and the identity M(k^2)=∫R(t)e^{ikt}dt (Eq. (16)) is a standard Fourier correspondence, not a fitted parameter renamed as a prediction. Controllability time in Theorem 3 is quoted from [29]; although one author is a coauthor of that monograph, the controllability result is an independent, published theorem, not the inverse uniqueness being proved. The leaf-peeling equations (44)–(48) algebraically compute the new reduced TW matrix from quantities already recovered, so the recursive step is constructive. Two flagged caveats belong to correctness rather than circularity. (1) Section 4, Step 4, Eq. (36): from C_i = S_in B_n S_in^{-1} only tr C_i and det C_i are used, giving the unordered pair {1/√ρ_n(0), 1/√μ_n(0)}; without an ordering or genericity assumption, swapping the two densities on the unmeasured root edge and rotating the local frame by 90° yields identical reduced boundary data, so the stated uniqueness in Theorems 1–2 is stronger than the proof supports. (2) Theorem 3 is explicitly only sketched ('We give a sketch of the proof; a detailed investigation ... will be presented in a separate paper'), so the dynamical result is incomplete as written. Neither of these is a self-justifying reduction; hence the circularity score is 1.
Assumptions & free parameters
assumptions (4)
- standard math The scalar BC method on an interval recovers rho(x) from the diagonal response R_ii(t), 0 < t < 2L_i.
- ad hoc to paper The transmitted first arrival between two boundary edges sharing a vertex is nonzero in the (1,1) channel at t = L_i + L_j, so Eq. (41) detects common vertices.
- ad hoc to paper The matrices A_i and C_i have distinct eigenvalues and the eigenvalue ordering identifies the u and w channels of unmeasured edges, so Eqs. (34)-(36) determine angles uniquely.
- domain assumption Each edge's local coordinate system is taken to be its tangent/normal frame, and this frame is part of the unknown to be recovered; for internal edges the data are invariant under a 90-degree rotation of the frame together with a swap of rho and mu.
Cite this review
Pith. "Pith review of On the inverse problem of the two-velocity tree-like graph." pith.science (2026). https://pith.science/paper/2MUKHGO5
@misc{pith2026250522138,
author = {Pith},
title = {Pith review of: On the inverse problem of the two-velocity tree-like graph},
year = {2026},
howpublished = {\url{https://pith.science/paper/2MUKHGO5}},
note = {Machine review of arXiv:2505.22138}
}
read the original abstract
In this article the authors continue the discussion in \cite{ALM} about inverse problems for second order elliptic and hyperbolic equations on metric trees from boundary measurements. In the present paper we prove the identifiability of varying densities of a planar tree-like network of strings along with the complete information on the graph, i.e. the lengths of the edges, the edge degrees and the angles between neighbouring edges. The results are achieved using the Titchmarch-Weyl function for the spectral problem and the Steklov-Poincar{\'e} operator for the dynamic wave equation on the tree. The general result is obtained by a peeling argument which reduces the inverse problem layer-by-layer from the leaves to the clamped root of the tree.
Reference graph
Works this paper leans on
-
[1]
Avdonin, Control problems on quantum graphs, in: Analysis on Graphs and its Applications, P
S. Avdonin, Control problems on quantum graphs, in: Analysis on Graphs and its Applications, P. Exner, J. Keating, P. Kuchment, T. Sunada, A. Teplyaev (eds.), Proceedings of Symposia in Pure Mathematics, AMS77, 507–521 (2008)
work page 2008
-
[2]
S. Avdonin, B. Belinskiy and M. Matthews, Dynamical inverse problems on a metric tree, Inverse Problems27(7), 1–21 (2011)
work page 2011
-
[3]
S.A. Avdonin, M.I. Belishev, and Yu.S. Rozhkov, The BC–method in the inverse problem for the heat equation, J. Inverse and Ill-Posed Problems5, 309–322 (1997)
work page 1997
-
[4]
S. Avdonin and J. Bell, Determining a distributed parameter in a neural cable model via a boundary control method, J. Mathematical Biology67(1), 123–141 (2013)
work page 2013
-
[5]
S. Avdonin and J. Bell, Determining a Distributed Conductance Parameter for a Neuronal Cable Model Defined on a Tree Graph, Journal of Inverse Problems and Imaging, submitted
-
[6]
S.A. Avdonin and S.A. Ivanov, Families of Exponentials. The Method of Moments in Controllability Problems for Distributed Parameter Systems (Cambridge Univ. Press., 1995)
work page 1995
-
[7]
S. A. Avdonin and P. B. Kurasov, Inverse problems for quantum trees, Inverse Probl. Imag.2(1), 1–21 (2008)
work page 2008
-
[8]
S. Avdonin, S. Lenhart, and V. Protopopescu, Solving the dynamical inverse problem for the Schr¨ odinger equation by the Boundary Control method, Inverse Problems18, 41–57 (2002)
work page 2002
Show all 33 references
-
[9]
Avdonin, G
S. Avdonin, G. Leugering and V. Mikhaylov, On an inverse problem for tree-like networks of elastic strings, Zeit. Angew. Math. Mech.90, 136–150 (2010)
2010
-
[10]
Avdonin and V
S. Avdonin and V. Mikhaylov, Controllability of partial differential equations on graphs, Appl. Math.35, 379–393 (2008)
2008
-
[11]
Avdonin and V
S. Avdonin and V. Mikhaylov, The boundary control approach to inverse spectral theory, Inverse Problems26(4), 1–19 (2010)
2010
-
[12]
S. A. Avdonin, V. S. Mikhaylov and A. V. Rybkin, The boundary control approach to the Titchmarsh–Weylm−function, Comm. Math. Phys.275(3), 791–803 (2007). 16 S. A VDONIN, A. CHOQUE RIVERO, G. LEUGERING, AND V. MIKHAYLOV
2007
-
[13]
Petersburg Math
Belishev M.I., Canonical model of a dynamical system with boundary control in inverse problem for the heat equation, St. Petersburg Math. Journal7(6), 869–890 (1996)
1996
-
[14]
M. I. Belishev, Recent progress in the boundary control method. Inverse Problems 23(5), R1–R67 (2007)
2007
-
[15]
M. I. Belishev, Boundary spectral inverse problem on a class of graphs (trees) by the BC method, Inverse Problems20, 647–672 (2004)
2004
-
[16]
M. I. Belishev, Boundary control and inverse problems: one-dimensional variant of BC-method, Zap. Nauchn. Sem. S.-Peterburg. Otdel. Mat. Inst. Steklov. (POMI)354 (2008), Mat. Vopr. Teor. Rasprostr. Vol. 37, 19–80; translation in J. Math. Sci. (N. Y.)155(3), 343–378 (2008)
2008
-
[17]
M. I. Belishev and V. S. Mikhaylov, Unified approach to classical equations of inverse problem theory, J. Inverse Ill-Posed Probl.20(4), 461–488 (2012)
2012
-
[18]
M. I. Belishev and A. F. Vakulenko, Inverse problems on graphs: recovering the tree of strings by the BC-method, Inverse Ill-posed Probl. Ser.14, 29–46 (2006)
2006
-
[19]
B. M. Brown and R. Weikard, A Borg-Levinson theorem for trees, Proc. R. Soc. Lond. Ser. A Math. Phys. Eng. Sci.,461(2062), 3231–3243 (2005)
2005
-
[20]
D´ ager and E
R. D´ ager and E. Zuazua, Wave propagation, observation and control in 1-dflexible multi-structures. Series: Math´ ematiques & Applications 50. Springer-Verlag, Berlin, 2006
2006
-
[21]
Freiling and V
G. Freiling and V. Yurko, Inverse problems for Strum-Liouville operators on noncom- pact trees, Results Math.50(3–4), 195–212 (2007)
2007
-
[22]
Kachalov, Y
A. Kachalov, Y. Kurylev, M. Lassas and N. Mandache, Equivalence of time-domain inverse problems and boundary spectral problems, Inverse Problems20, 491–436 (2004)
2004
-
[23]
Kostrykin and R
V. Kostrykin and R. Schrader, Kirchhoff’s rule for quantum wires, J. Phys A: Math. Gen.32, 595–630 (1999)
1999
-
[24]
Kuchment, Quantum graphs: an introduction and a brief survey, in: Analysis on Graphs and its Applications, P
P. Kuchment, Quantum graphs: an introduction and a brief survey, in: Analysis on Graphs and its Applications, P. Exner, J. Keating, P. Kuchment, T. Sunada, A. Teplyaev (eds.), Proceedings of Symposia in Pure Mathematics, AMS77, 219–314 (2008)
2008
-
[25]
Kurasov and F
P. Kurasov and F. Stenberg, On the inverse scattering problem on branching graphs, J. Phys. A: Math. Gen.35, 101–121 (2002)
2002
-
[26]
J. E. Lagnese and G. Leugering, Domain decomposition methods in optimal control of partial differential equations, ISNM. International Series of Numerical Mathematics 148, Basel: Birkh¨ auser, 2004
2004
-
[27]
J. E. Lagnese, G. Leugering and E. J. P. G. Schmidt, Control of planar networks of Timoshenko beams. SIAM J. Control Optimization31(3), 780–811 (1993)
1993
-
[28]
J. E. Lagnese, G. Leugering and E. J. P. G. Schmidt, On the analysis and control of hyperbolic systems associated with vibrating networks. Proc. R. Soc. Edinb., Sect. A 124(1), 77–104 (1994)
1994
-
[29]
J. E. Lagnese, G. Leugering and E. J. P. G Schmidt, Modeling, analysis and control of dynamic elastic multi-link structures, Birkh¨ auser Boston, Systems and Control: Foundations and Applications, 1994
1994
-
[30]
Leugering, and E
G. Leugering, and E. J. P. G. Schmidt, On exact controllability of networks of non- linear elastic strings in 3–dimensional space, Chin. Ann. Math. Ser. B 33(1), 33–60 (2012)
2012
-
[31]
Leugering and J
G. Leugering and J. Sokolowski, Topological sensitivity analysis for elliptic problem on graphs, Control and Cybernetics37(4), 971–997 (2008)
2008
-
[32]
Leugering and E
G. Leugering and E. Zuazua, On exact controllability of generic trees, ESAIM, Proc. 8, 95–105 (2000)
2000
-
[33]
Yurko, Inverse Sturm-Liouville operator on graphs, Inverse Problems21, 1075– 1086 (2005)
V. Yurko, Inverse Sturm-Liouville operator on graphs, Inverse Problems21, 1075– 1086 (2005). TWO-VELOCITY TREE-LIKE GRAPH 17 Department of Mathematics and Statistics, University of Alaska, F air- banks, AK 99775-6660, USA. Instituto de F´ısica y Matem´aticas, Universidad Micho...
2005
Reviewed August 7, 2026 · model on record in the stance chip above.
Discussion (0). Sign in to comment.