REVIEW 2 major objections 4 minor 30 references
Finding hypergraph immersion is fixed-parameter tractable
T0 review · 2 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash
Pith's one-line read Hypergraph immersion is decidable in $O(|V(G)|^6)$ time for every fixed pattern.
desk verdict The hypergraph immersion definition is a real contribution, but the claimed O(N^6) FPT algorithm rests on a false hypergraph Mader lemma and an out-of-scope invocation of Grohe et al.; the main theorem is not established. 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 paper gives two equivalent definitions of hypergraph immersion: a geometric definition using edge-disjoint connected subgraphs (hypergraph paths known as Berge paths), and an operation-based definition using two rewrites—edge coalescence, which merges two intersecting hyperedges into their union, and dewetting, which removes one vertex from a hyperedge. The algorithmic proof is carried by three constructions: the $M$-generalised factor graph $G'_M$, an ordinary graph with $M$ copies of each vertex and one vertex per hyperedge; densification, which replaces each distinguished vertex copy by a large clique $K_L$ to enforce the correct vertex mapping; and divisions of $H$, which replace each hyperedge by a Steiner tree connecting its vertices so that all possible internal topologies of an immersed hyperedge fall into finitely many topological classes. The densified factor graphs turn hypergraph immersion into ordinary graph embedding, for which a cubic-time algorithm is available, and the final complexity is obtained by bounding the size of the auxiliary graph.
What would settle it
Test Lemma 4.5 directly: take $H = K_2$ (a single edge) and take $G$ to be a hypergraph on $n$ vertices that consists only of $n$ singleton hyperedges, one at each vertex, so $|E(G)| = |V(G)|$. By the definition of hypergraph paths, no Berge path connects two distinct vertices in such a $G$, so $H$ cannot be immersed; this would disprove the lemma as stated and show that the proof's linear edge cap needs a different justification.
Extended reading notes
Core claim
The central claim, Theorem 5.1 (with Theorem 3.5 as the general statement), is that for every finite loopless hypergraph $H$ there is an $O(|V(G)|^6)$-time algorithm deciding whether $H$ can be immersed in a hypergraph $G$. Hypergraph immersion means mapping the vertices of $H$ injectively to vertices of $G$ and mapping each hyperedge of $H$ to an edge-disjoint connected subgraph of $G$ containing the images of its vertices. The proof first handles ordinary graphs $H$, using an $M$-generalised factor graph $G'_M$ and a densification step that replaces special vertices by large cliques so that the vertex mapping is forced to respect the original hypergraph vertices; it then handles hypergraph $H$ by testing factor graphs of a finite set of divisions of $H$, which account for all possible internal topologies of an immersed hyperedge. A hypergraph analogue of the classical average-degree forcing theorem is used to bound $|E(G)|$ linearly in $|V(G)|$, which keeps the auxiliary graph at size $O(|V(G)|^2)$ and gives the stated $O(|V(G)|^6)$ bound.
Load-bearing premise
The proof's $O(|V(G)|^6)$ bound depends on Lemma 4.5, which asserts that any hypergraph with at least a constant $C(H)$ times $|V(G)|$ edges already contains $H$ as an immersion; the lemma is proved by reducing the hypergraph to an ordinary graph through dewetting and capping multiedges, and if that step fails, the linear edge bound on $|E(G)|$ and hence the stated exponent are not established.
Editorial extensions
If this is right
- For any fixed pattern $H$, deciding immersion in a hypergraph $G$ takes at most $O(|V(G)|^6)$ time, so the problem is fixed-parameter tractable with exponent independent of $H$.
- The ordinary graph immersion problem is the special case where every hyperedge has size 2, so the theorem supplies a polynomial immersion test for ordinary graphs as well, with a larger but still constant exponent.
- The finite division set $D(H)$ means the search over possible internal topologies of immersed hyperedges is finite and depends only on the pattern $H$, not on the input hypergraph.
- By the paper's duality theorem, the same $O(|V(G)|^6)$ algorithm also decides dual immersion between the transposed hypergraphs, giving a polynomial check for the vertex-coalescence version of the problem.
Reading between the lines
- Editorial inference: the constant hidden by the $O(|V(G)|^6)$ bound grows with the number of divisions in $D(H)$—the paper itself counts 18 topological classes for the complete 3-uniform hypergraph on 4 vertices—so the practical bottleneck is pattern-dependent even though the asymptotic exponent is not.
- Editorial inference: the same reduction to ordinary graph embedding could be reused for other hypergraph containment relations that allow arbitrary internal topology for each hyperedge, since the division-and-densification machinery is not specific to immersion.
- Editorial inference: the exact exponent 6 is tied to the linear edge bound supplied by the hypergraph analogue of the average-degree forcing theorem; if that bound needs to be weakened, the paper's method would still give a polynomial algorithm, but possibly with a larger fixed exponent.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper defines a notion of hypergraph immersion, in which each hyperedge of a pattern H is mapped to an edge-disjoint connected subgraph (with respect to Berge paths) of a host hypergraph G, and it proposes an equivalent characterization via two operations, coalescence and dewetting. The main claim (Theorems 3.5, 4.1, and 5.1) is that for every fixed ordinary or hypergraph H, deciding whether H can be immersed in a hypergraph G can be done in O(|V(G)|^6) time, giving fixed-parameter tractability. The proof strategy is to convert hypergraph immersion to ordinary graph embedding by using an M-generalized factor graph, densifying vertices with large cliques, and, for hypergraph patterns, testing a finite set of 'divisions' of H. The paper also introduces a dual hypergraph immersion notion via transposition. The algorithmic argument is not established as written: a key lemma bounding the number of hyperedges is false, and the densified pattern used in the reduction is not fixed.
Significance. The paper proposes a natural and useful extension of graph immersion to hypergraphs, and the division-based reduction is conceptually interesting. If the technical gaps were repaired, an O(|V(G)|^6) fixed-parameter algorithm for hypergraph immersion would be a meaningful contribution to algorithmic graph theory and could be relevant to quantum network routing. The paper is also clearly written and the definitions are well motivated. However, the central theorem is not proven in the current version because of load-bearing errors in the complexity analysis, so the significance cannot be assessed as claimed.
major comments (2)
- [§4.2, Lemma 4.5] The lemma is false as stated. The paper permits size-1 hyperedges (see the footnote in Section 2.2 and Remark 2.3). Let G have n vertices and C·n distinct singleton hyperedges, where C is any proposed constant. Then |E(G)| = C|V(G)|, but G contains no Berge path between distinct vertices, so the ordinary graph K2 cannot be immersed in G. Thus the claimed bound |E(G)| = O(|V(G)|) is not a consequence of the lemma. The proof's reduction via dewetting is also invalid: applying dewetting to a hyperedge e yields the smaller hyperedge e\{v}, and for |e|=1 it cannot produce an ordinary edge at all; it never turns a hyperedge into two ordinary edges. Consequently, the step in the proof of Theorem 4.1 that uses Lemma 4.5 to conclude |V(G''_{M,L})| = O(|V(G)|^2) fails, and the O(|V(G)|^6) bound is not established.
- [§4.2, Remark 4.4 and proof of Theorem 4.1] The densified pattern H''_L is not a fixed graph. Remark 4.4 sets L > M|V(G'_M)|, so even if Lemma 4.5 were true, |V(G'_M)| = M|V(G)| + |E(G)| = O(|V(G)|), giving |V(H''_L)| = L|V(H)| = O(|V(G)|). Theorem 2.7 provides an O(|V(G)|^3) algorithm for embedding a fixed graph H; the constant in the O may depend on H. Applying Theorem 2.7 to the growing graph H''_L therefore does not yield a uniform O(|V(G)|^6) algorithm. The same problem appears in the proof of Theorem 5.1, where each densified division \tilde H''_L also has size depending on L = O(|V(G)|). The paper acknowledges the growth of L in Remark 4.4 but gives no argument that the dependence on H''_L is harmless for fixed-parameter tractability.
minor comments (4)
- [§3.1, Proposition 3.4] The proof of Proposition 3.4 is only a sketch and relies on the authors' own preprint [28]. In the 'if' direction, it is not shown rigorously how a sequence of coalescence and dewetting operations on a subgraph of G yields well-defined edge-disjoint connected subgraphs α(e) for each e ∈ E(H). An independent, complete proof would be preferable, or the statement should be clearly marked as a result from [28] with a full proof.
- [§5.4, proof of Theorem 5.1] The running time for testing one densified factor graph is stated as O(|V(G''_{M,L})|) (twice in the proof of Theorem 5.1); it should be O(|V(G''_{M,L})|^3) to match Theorem 2.7.
- [§4.2, Lemma 4.5] The statement 'Let |E(G)| = C |V(G)|' should be 'Let |E(G)| ≥ C |V(G)|' for the lemma to be a Mader-type density condition. Also, the expression '|E(G0)| / |E(H)| = C0 |V(G0)|' is dimensionally confusing and should be rewritten with explicit multiplicative constants.
- [§3.1, proof of Proposition 3.4] There is a typographical error: 'a(v1) ≠ α(v2)' should read 'α(v1) ≠ α(v2)'.
Circularity Check
No circular derivation found: the main reductions unpack Definition 3.1 rather than presupposing the target; the serious flaw in Lemma 4.5 is an unsupported (and false) mathematical claim, not a circularity.
full rationale
The paper does not fit parameters to data, rename a known result, or define its target in terms of itself. Its central reduction chain goes: hypergraph immersion (Definition 3.1) -> M-generalized factor graph (Section 4.1) -> densified graph embedding (Lemma 4.3) -> the external embedding algorithm of Grohe et al. (Theorem 2.7), with the edge-count bound supplied by Lemma 4.5. Each step either unpacks the definitions or cites an external, independently published theorem. The only self-citation is Proposition 3.4, credited to the authors' Ref. [28], but the paper also gives a proof sketch in the text, and later sections do not use the coalescence/dewetting equivalence as a load-bearing step; they work directly from Definition 3.1. Under the stated rules, this is at most a minor self-citation, not circularity. The most serious problem is correctness, not circularity: Lemma 4.5's proof claims that dewetting reduces any hypergraph to an ordinary graph with 'the same number of vertices and (ordinary) edges,' but dewetting a hyperedge of size greater than two does not produce pairwise ordinary edges while preserving the connectivity needed for an immersion. The lemma is therefore false (e.g., linearly many singleton hyperedges give no Berge path), but a false or insufficiently proved lemma is an unsupported assumption, not an equivalence-by-construction between an input and a prediction. For the same reason, the potential issue that Theorem 2.7 is applied to a densified graph whose size grows with L falls outside the fixed-H hypothesis; that is a gap in the argument, not a circular step. No quoted equation or construction in the paper reduces to its own input, so the circularity score is 0.
Assumptions & free parameters
free parameters (3)
- M (number of vertex copies in M-generalized factor graph) =
M > |E(H)|, chosen as O(1)
- L (clique size in densification) =
L > M|V(G'_M)|, so L = O(|V(G)|) in the sparse case
- C(H) (Mader constant for hypergraphs) =
not quantified
assumptions (4)
- standard math Mader's theorem for ordinary graphs
- standard math Grohe et al. Theorem 2.7: fixed-parameter tractability of graph embedding
- domain assumption Definition of hypergraph immersion (Definition 3.1) and equivalence with coalescence and dewetting (Proposition 3.4)
- ad hoc to paper Lemma 4.5: hypergraph Mader bound
invented entities (2)
-
M-generalized factor graph G'_M
-
Divisions of hypergraphs and topological equivalence classes (D(H))
Cite this review
Pith. "Pith review of Finding hypergraph immersion is fixed-parameter tractable." pith.science (2026). https://pith.science/paper/PTHXK3N4
@misc{pith2026241116017,
author = {Pith},
title = {Pith review of: Finding hypergraph immersion is fixed-parameter tractable},
year = {2026},
howpublished = {\url{https://pith.science/paper/PTHXK3N4}},
note = {Machine review of arXiv:2411.16017}
}
abstract
Immersion minor is an important variant of graph minor, defined through an injective mapping from vertices in a smaller graph $H$ to vertices in a larger graph $G$ where adjacent elements of the former are connected in the latter by edge-disjoint paths. Here, we consider the immersion problem in the emerging field of hypergraphs. We first define hypergraph immersion by extending the injective mapping to hypergraphs. We then prove that finding a hypergraph immersion is fixed-parameter tractable, namely, there exists an $O(N^6)$ polynomial-time algorithm to determine whether a fixed hypergraph $H$ can be immersed in a hypergraph $G$ with $N$ vertices. Additionally, we present the dual hypergraph immersion problem and provide further characteristics of the algorithmic complexity.
Figures
Figures from the paper (4 more)
Reference graph
Works this paper leans on
-
[28]
Y. Tian, Y. Liu, and X. Meng, Quantum networks: from multipartite entanglement to hyper- graph immersion, preprint (2024), arXiv:2406.13452
work page Pith review arXiv 2024
-
[1]
Claude Berge, Graphs and hypergraphs, North-Holland, Amsterdam, Netherlands, 1973
work page 1973
-
[2]
G. Bianconi, Higher-order networks , Elements in the Structure and Dynamics of Complex Networks, Cambridge University Press, 2021
work page 2021
-
[3]
Dorogovtsev, The theory of percolation on hypergraphs , arXiv preprint (2023), arXiv:2305.12297
Ginestra Bianconi and Sergey N. Dorogovtsev, The theory of percolation on hypergraphs , arXiv preprint (2023), arXiv:2305.12297
arXiv 2023
-
[4]
C. Buchheim, M. Chimani, C. Gutwenger, M. J¨ unger, and P. Mutzel, Crossings and pla- narization, pp. 43–79, CRC Press, 2014
work page 2014
-
[5]
Patricia Contreras-Tejada, Carlos Palazuelos, and Julio I. de Vicente, Genuine Multipartite Nonlocality Is Intrinsic to Quantum Networks , Physical Review Letters 126 (2021), no. 4, 040501
work page 2021
-
[6]
R.G. Downey and M.R. Fellows, Fixed-parameter intractability, [1992] Proceedings of the Seventh Annual Structure in Complexity Theory Conference, 1992, pp. 36–49
work page 1992
-
[7]
M. Fellows and M. Langston, Nonconstructive tools for proving polynomial-time decidability , J. ACM 35 (1988), no. 3, 727–739
work page 1988
Show all 30 references
-
[8]
Fellows and M
M. Fellows and M. Langston, On Well-Partial-Order Theory and Its Application to Combi- natorial Problems of VLSI Design , SIAM J. Discrete Math. 5 (1992), no. 1, 117–126
1992
-
[9]
Martin Grohe, Ken-ichi Kawarabayashi, D´ aniel Marx, and Paul Wollan, Finding topological subgraphs is fixed-parameter tractable, Proceedings of the Forty-Third Annual ACM Sympo- sium on Theory of Computing (New York, NY, USA), STOC ’11, Association for Computing Machinery, 20...
2011
-
[10]
Gross and Jay Yellen, Graph theory and its applications , second ed., CRC Press, Boca Raton, FL, USA, 2005
Jonathan L. Gross and Jay Yellen, Graph theory and its applications , second ed., CRC Press, Boca Raton, FL, USA, 2005
2005
-
[11]
F. Hahn, A. Pappa, and J. Eisert, Quantum network routing and local complementation , npj Quantum Information 5 (2019), no. 1, 1–7
2019
-
[12]
Jost and R
J. Jost and R. Mulas, Hypergraph laplace operators for chemical reaction networks , Adv. Math. 351 (2019), 870–896
2019
-
[13]
Klamt, U
S. Klamt, U. Haus, and F. Theis, Hypergraphs and cellular networks , PLOS Computat. Biol. 5 (2009), no. 5, 1–6
2009
-
[14]
Krumov, C
L. Krumov, C. Fretter, M. M¨ uller-Hannemann, K. Weihe, and M. H¨ utt, Motifs in co- authorship networks and their relation to the impact of scientific publications , Eur. Phys. J. B 84 (2011), no. 1, 535–540
2011
-
[15]
J. Lee, K. I. Goh, D. S. Lee, and B. Kahng, (k, q)-core decomposition of hypergraphs, Chaos Solitons Fractals 173 (2023), 113645
2023
-
[16]
Munro, Aggregating quantum networks, Physical Review A 102 (2020), no
Nicolo Lo Piparo, Michael Hanks, Kae Nemoto, and William J. Munro, Aggregating quantum networks, Physical Review A 102 (2020), no. 5, 052613
2020
-
[17]
Lov´ asz,Graph minor theory , Bulletin of the American Mathematical Society 43 (2005), no
L. Lov´ asz,Graph minor theory , Bulletin of the American Mathematical Society 43 (2005), no. 1, 75–86
2005
-
[18]
Mader, Homomorphieeigenschaften und mittlere Kantendichte von Graphen
W. Mader, Homomorphieeigenschaften und mittlere Kantendichte von Graphen. , Mathema- tische Annalen 174 (1967), 265–268. 20 XIANGYI MENG AND YU TIAN
1967
-
[19]
thesis, Sor- bonne Universit´ e, 2021
Cl´ ement Meignant,Multipartite communications over quantum networks , Ph.D. thesis, Sor- bonne Universit´ e, 2021
2021
-
[20]
Kov´ acs,Path Percolation in Quantum Communication Networks, arXiv preprint (2024), arXiv:2406.12228
Xiangyi Meng, Bingjie Hao, Bal´ azs R´ ath, and Istv´ an A. Kov´ acs,Path Percolation in Quantum Communication Networks, arXiv preprint (2024), arXiv:2406.12228
2024 arXiv
-
[21]
Kov´ acs,Quantum Networks Enhanced by Distributed Quantum Memories , preprint (2024), arXiv:2403.16367
Xiangyi Meng, Nicol` o Lo Piparo, Kae Nemoto, and Istv´ an A. Kov´ acs,Quantum Networks Enhanced by Distributed Quantum Memories , preprint (2024), arXiv:2403.16367
2024
-
[22]
Mihir Pant, Hari Krovi, Don Towsley, Leandros Tassiulas, Liang Jiang, Prithwish Basu, Dirk Englund, and Saikat Guha, Routing entanglement in the quantum internet , npj Quantum Information 5 (2019), no. 1, 25
2019
-
[23]
Stefano Pirandola, End-to-end capacities of a quantum communication network , Communi- cations Physics 2 (2019), no. 1, 1–10
2019
-
[24]
Robertson and P
N. Robertson and P. Seymour, Graph minors .xiii. the disjoint paths problem , Journal of Combinatorial Theory, Series B 63 (1995), no. 1, 65–110
1995
-
[25]
Nash-Williams’ immersion conjec- ture, Journal of Combinatorial Theory, Series B 100 (2010), no
Neil Robertson and Paul Seymour, Graph minors XXIII. Nash-Williams’ immersion conjec- ture, Journal of Combinatorial Theory, Series B 100 (2010), no. 2, 181–205
2010
-
[26]
Evan Sutcliffe and Alejandra Beghelli, Multiuser Entanglement Distribution in Quantum Net- works Using Multipath Routing , IEEE Transactions on Quantum Engineering 4 (2023), 1–15
2023
-
[27]
Taramasco, Cointet J., and C
C. Taramasco, Cointet J., and C. Roth, Academic team formation as evolving hypergraphs , Scientometrics 85 (2010), no. 1, 721–740
2010
-
[29]
Wagner, ¨Uber eine eigenschaft der ebenen komplexe , Math
K. Wagner, ¨Uber eine eigenschaft der ebenen komplexe , Math. Ann. 114 (1937), 570–590
1937
-
[30]
26, 260506
Meihong Wang, Yu Xiang, Haijun Kang, Dongmei Han, Yang Liu, Qiongyi He, Qihuang Gong, Xiaolong Su, and Kunchi Peng, Deterministic Distribution of Multipartite Entanglement and Steering in a Quantum Network by Separable States , Physical Review Letters 125 (2020), no. 26, 26050...
2020
Reviewed August 12, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.