Pith. sign in

REVIEW 3 major objections 6 minor 42 references

Decategorification turns DAG magnitude-homology Euler characteristics into the coefficients of an inverse q-matrix, computed by polynomial back-substitution.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-07-31 23:38 UTC pith:EN4RNGBN

load-bearing objection Solid decategorification-based computation for magnitude homology Euler characteristics of DAGs; the MLP application is a suggestive but uncontrolled proof of concept. the 3 major comments →

arxiv 2607.23357 v1 pith:EN4RNGBN submitted 2026-07-25 math.AT math.CO

Magnitude homology and Euler characteristics of directed acyclic graphs

Steve Huntsman This is my paper
classification math.AT math.CO MSC 05C2055N3555U1068R10
keywords magnitude homologyEuler characteristicdirected acyclic graphsdecategorificationlayered DAGsmultilayer perceptronsMöbius functionorder complex
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Directed acyclic graphs carry a homology theory called magnitude homology, whose Euler characteristic summarizes path structure. The paper's central claim is that for a DAG with one source and one target, this Euler characteristic can be read off from the inverse of a q-weighted distance matrix, without ever building the homology. That reduces the computation to exact integer polynomial back-substitution, which the paper implements and tests; it also proves structural facts about layered DAGs and about gluing DAGs together. The same technique is then applied to fixed-size graphs extracted from trained multilayer perceptrons, yielding Euler characteristic patterns that separate handwritten-digit classes even when vertex and arc counts are held constant. A sympathetic reader would care because this makes a previously expensive invariant cheap and usable as a network fingerprint.

Core claim

The central claim is that decategorification turns a DAG's magnitude homology Euler characteristic into linear algebra: (Z_q^{-1})_{st} = \sum_L \chi^{(s,t)}_{*,L} q^L, with (Z_q)_{uv}=q^{d(u,v)}. The paper computes these coefficients exactly by polynomial back-substitution, with a floating-point alternative that it warns can fail at roughly a thousand vertices. It proves that in layered DAGs the chain complex is an order complex, so Euler characteristics become Möbius values, and that gluing two st-DAGs at one vertex kills all source-to-target magnitude homology. Its MLP proof of concept uses fixed-size st-DAGs and reports class-discriminative Euler characteristic grids.

What carries the argument

The q-zeta matrix Z_q, whose (u,v) entry is q^{d(u,v)} when v is reachable from u and zero otherwise. Its (s,t) inverse entry is the generating function over path lengths of the Euler characteristics, turning the computation into polynomial back-substitution in reverse topological order. The second load-bearing object is the isomorphism, for layered DAGs, between the magnitude chain complex at the natural path length and the augmented simplicial chain complex of the open interval order complex.

Load-bearing premise

The load-bearing premise for the neural-network application is that the hand-chosen node and arc scores used to select which neurons and arcs form each graph truly reflect the network's activation geometry; there is no comparison against alternative scoring rules, and the mathematical part separately assumes the standard decategorification identity.

What would settle it

Directly compute the magnitude Betti numbers for a small layered st-DAG, such as one with layer widths 5,4,3,2, and check that their alternating sums agree with the coefficients from the q-zeta inverse. Separately, rerun the MLP analysis with permuted node and arc scores while keeping the DAG sizes fixed: if the reported class structure persists, the discriminative signal is not coming from those score definitions.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • Euler characteristics for st-DAGs become computable by exact integer polynomial arithmetic without ever constructing the magnitude chain complex.
  • For layered DAGs, the source-to-target Euler characteristic at the natural path length equals the Möbius value, so existing fast Möbius-transform algorithms can be reused.
  • Gluing two st-DAGs at a single vertex produces zero source-to-target magnitude homology at every path length, making single-vertex interfaces invisible to this invariant.
  • The MLP construction holds the number of vertices and arcs per layer fixed, so any class-discriminative pattern in the Euler characteristics is not attributable to coarse graph size.
  • The appendix gives two short implementations, one fast floating-point and one exact polynomial; for graphs with roughly a thousand vertices the paper recommends trusting the exact version.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Editorial extension: the same back-substitution could be applied to all source-target pairs at once, yielding a matrix of Euler characteristics that would serve as a much richer graph invariant than a single st entry.
  • Editorial extension: the gluing theorem suggests a width interpretation—replacing the single identified vertex by a set of interface vertices should partially undo the killing, and the rank of surviving homology could quantify information flow across the interface.
  • Editorial extension: the MLP result could be stress-tested by replacing the activation-based scores with random scores on the same fixed-size DAGs; if class separation persists, the visible structure is a property of the fixed architecture rather than of the scoring rule.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper develops a method for computing the magnitude-homology Euler characteristic of a directed acyclic graph (DAG) via the decategorification identity (exp[-τd])^{-1}_{st} = Σ_L χ^{(s,t)}_{•,L} exp(-τL). For st-DAGs, the authors propose exact polynomial back-substitution on the q-zeta matrix and a floating-point heuristic solver, both given in the appendix. The mathematical core includes Theorem 2 (vanishing of magnitude homology for a DAG obtained by identifying the target of one st-DAG with the source of another), Lemma 4 and Corollaries 5–6 (closed-form Betti numbers for layered complete MLP DAGs), and Theorem 7 with Corollary 8 (identification of the magnitude chain complex with the order complex for layered DAGs, yielding a Möbius-function formula). As a proof of concept, Section 6.3 constructs st-DAGs from hand-designed node and arc scores derived from ReLU MLP activations on MNIST and claims that the resulting Euler characteristics reveal class-discriminative structure.

Significance. The theoretical results are sound and the exact polynomial algorithm in §A.2 is a practical, reproducible method for computing magnitude-homology Euler characteristics without constructing the chain complex. The identification with the order complex for layered DAGs (Theorem 7) and the Möbius-function consequence (Corollary 8) are clean and potentially useful. If the MLP application were properly validated, it would be an interesting new tool for neural-network analysis; however, the current evidence is anecdotal and lacks the controls needed to attribute the observed structure to the Euler characteristic itself.

major comments (3)
  1. [§6.3 and Abstract] The claim that the Euler characteristic 'recovers class-discriminative structure' is not supported by the presented evidence. The st-DAGs are not derived from the network alone; they are selected using the class-conditional scores ν_{s,t,X}(u) and α_{s,t,X}(u,u') defined in §6.3, which depend directly on the true-class subset C_t and on input-pixel magnitudes |a_s(x)|. Figures 3–4 are qualitative heatmaps, with no null model, no ablation, and no quantitative separation metric. A control experiment using class-agnostic or random scores (while keeping k and the arc budget fixed) is essential to establish that the discriminative signal comes from the magnitude-homology Euler characteristic rather than from the hand-designed score construction itself.
  2. [§5 and §A.2] The abstract's 'scalable' claim is not benchmarked. The paper states in §5 that for large instances 'on the order of a thousand or more vertices' the floating-point method sometimes disagrees with exact arithmetic, yet no asymptotic complexity or wall-clock measurements are reported for either algorithm. The exact polynomial back-substitution in §A.2 appears to be O(n^3) in the worst case, but this is not stated. Given that the MLP demo uses graphs with only a few hundred vertices, the scalability claim requires explicit complexity analysis and experiments on larger DAGs to be credible.
  3. [§5, Eq. (7)–(8)] The algorithmic exposition over-emphasizes the floating-point Vandermonde/Babai heuristic, which is explicitly described as numerically unreliable at scale, while the exact polynomial method is relegated to an appendix. Since the polynomial method is both exact and simple, the paper should present it as the primary algorithm, state its complexity, and clarify the practical regime (if any) in which the floating-point method is preferable. The current organization undercuts the paper's own 'scalable' claim.
minor comments (6)
  1. [§4, Lemma 4 proof] The proof of Lemma 4 shows that the displayed chains are cycles, but the assertion that 'any chain not in this space has nonzero boundary' is not demonstrated. This is likely fixable with a short argument, but as written the proof is incomplete.
  2. [§3, Eq. (1)] The direct sum decomposition (1) is cited to [23] but should state explicitly that the summands are subcomplexes generated by simplices with fixed initial and terminal vertices. The sentence currently follows the equation without a clear anchor.
  3. [§5] The phrase 'numerically truculent' is informal; recommend 'poorly conditioned' throughout. Also, the comment that integer least squares is NP-hard and the mention of 'closest vector problems' are tangential and can be omitted.
  4. [§6.3, Figures 3–4] The caption states that the grayscale is trimmed at extremes but does not specify the trimming rule (e.g., percentiles or absolute values). This should be quantified for reproducibility.
  5. [§6.3, last paragraph] The sentence 'the structure that is captured in the figures precisely encodes nonconvex activation geometry' overstates the evidence; 'appears to encode' would be more appropriate given the absence of controls.
  6. [§6.3, Definition 3] The notation K → n1,...,nM is typographically awkward and can be confused with the arrow in the definition. Consider using a different symbol, e.g., K(n1,...,nM).

Circularity Check

0 steps flagged

No significant circularity: the mathematical derivation rests on external decategorification and order-complex theorems, and the self-citations are not load-bearing.

full rationale

The central computation is not circular. The paper uses the standard decategorification identity (7) from Leinster-Shulman [13] to express the magnitude-homology Euler characteristic as coefficients of (Z_q^{-1})_{st}, and the polynomial back-substitution in §A.2 is an exact inversion of Z_q; interpreting the resulting coefficients as Euler characteristics is a theorem application, not a fit. Theorem 2, Lemma 4, Corollaries 5/6/8 and Proposition 9 are proved from the definitions or from external results ([15], [18], [19,20], [26]), with no parameter fitted to the target. Self-citations [28] and [42] are comparative/contextual and are not load-bearing: [28] merely supplies path-homology Betti numbers for comparison and a scalability remark, and [42] appears only in a closing remark about future work. The §6.3 MLP proof-of-concept builds st-DAGs using class-conditional node/arc scores ν_{s,t,X} and α_{s,t,X}, so the visual class-discriminative structure in Figures 3–4 could be carried by those hand-designed scores rather than by an independent property of the Euler characteristic; this is an application-level confound and an absence of an ablation/null model, not a circular derivation of χ from its own output. The mathematical claims are self-contained against external benchmarks, so no circular step is established.

Axiom & Free-Parameter Ledger

3 free parameters · 6 axioms · 0 invented entities

The mathematical core is self-contained against standard results: the decategorification identity and order-complex equivalences are cited, not introduced. The only genuinely ad hoc inputs are the algorithmic hyperparameters and the MLP score construction, which are application-level choices rather than new mathematical entities. No new particles, dimensions, or conserved quantities are postulated.

free parameters (3)
  • k (neurons retained per layer) = 64
    In §6.3, k=64 is chosen by hand to prune each hidden layer; it controls the size of the constructed st-DAGs and is not fitted, but the MLP results could depend on it.
  • arc budget (2k arcs per adjacent layer) = 128
    Also in §6.3, the number of arcs between adjacent hidden layers is set to 2k by construction; this is an ad hoc choice to hold graph structure fixed.
  • q-Ansatz in floating-point algorithm = q_i = (1+1/|L|)^i
    In §5 and §A.1, this choice is a heuristic to mitigate Vandermonde conditioning; the exact polynomial-arithmetic variant avoids it, so it is not load-bearing for the core claim.
axioms (6)
  • standard math Decategorification identity (7): (exp[-τ d])^{-1}_{st} = Σ (-1)^k β_{k,L} exp(-τL)
    Taken from Leinster-Shulman [13]; the entire computational method rests on this theorem. It is a known result, but the paper does not re-derive it.
  • standard math Philip Hall theorem: μ_hat_D = reduced Euler characteristic of order complex
    Used in Corollary 8 to identify layered-DAG Euler characteristics with Möbius numbers; standard combinatorial topology.
  • standard math Künneth theorem for simplicial joins
    Used in Proposition 9 to compute reduced Betti numbers of [n2]*...*[nM-1]; cited from Wachs [18].
  • domain assumption All digraphs are finite
    Stated in §2; finiteness is needed for the inverse zeta matrix to be a finite polynomial and for Betti numbers to be finite.
  • domain assumption Betti numbers are ranks over a field
    The paper does not specify the coefficient ring beyond the abstract chain complex; rank computations (e.g., Lemma 4) implicitly assume a field.
  • ad hoc to paper Node/arc score construction reflects activation geometry
    In §6.3, the specific definitions of ν_s,t,X and α_s,t,X are introduced without external validation; the MLP result claim depends on this selection being meaningful.

pith-pipeline@v1.3.0-alltime-deepseek · 11270 in / 14537 out tokens · 141081 ms · 2026-07-31T23:38:56.309949+00:00 · methodology

0 comments
read the original abstract

We develop a scalable approach to computing the magnitude homology Euler characteristic for directed acyclic graphs based on decategorification. Along with motivating mathematical results and some simple controlled examples, we deploy the Euler characteristic in a proof of concept application to the dynamic analysis of multilayer perceptrons, recovering class-discriminative structure while holding simpler subgraph properties fixed.

Figures

Figures reproduced from arXiv: 2607.23357 by Steve Huntsman.

Figure 1
Figure 1. Figure 1: χ ≡ χ (s,t) •,d(s,t) for layered st-DAGs with six vertices in each of three intermediate layers and eight arcs between intermediate layers [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Upper left: a histogram of χ ≡ χ (s,t) •,d(s,t) for randomly sampled layered st-DAGs with six vertices in each of three intermediate layers, and ten arcs between adjacent intermediate layers. Other panels: representative st-DAGs for indicated χ. fixed numbers of arcs per layer. This ensures that any features brought out by Euler characteristics are not attributable to any coarse graph structure. 7 The basi… view at source ↗
Figure 3
Figure 3. Figure 3: χ (s,t) •,d(s,t) for st-DAGs formed between input and output neurons as described in the main text. The underlying ReLU MLP has three hidden layers of width 128. Rows correspond data in true classes and columns correspond to data in predicted classes; blank entries correspond to zeros in the confusion matrix. Subpanels show the Euler characteristic per input neuron/pixel; all are on a common linear graysca… view at source ↗
Figure 4
Figure 4. Figure 4: As in [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

42 extracted references · 5 linked inside Pith

  1. [1]

    The magnitude of metric spaces.Documenta Mathematica, 18:857–905, 2013

    Tom Leinster. The magnitude of metric spaces.Documenta Mathematica, 18:857–905, 2013. 1

    2013

  2. [2]

    The magnitude of a metric space: from category theory to geometric measure theory

    Tom Leinster and Mark W Meckes. The magnitude of a metric space: from category theory to geometric measure theory. In Nicola Gigli, editor,Measure Theory in Non-Smooth Spaces, pages 156–193. De Gruyter, 2017

    2017

  3. [3]

    Cambridge, 2021

    Tom Leinster.Entropy and Diversity: the Axiomatic Approach. Cambridge, 2021. 1 9 Magnitude homology and Euler characteristics of DAGs

    2021

  4. [4]

    Weighting vectors for machine learning: numerical harmonic analysis applied to boundary detection.arXiv preprint arXiv:2106.00827, 2021

    Eric Bunch, Jeffery Kline, Daniel Dickinson, Suhaas Bhat, and Glenn Fung. Weighting vectors for machine learning: numerical harmonic analysis applied to boundary detection.arXiv preprint arXiv:2106.00827, 2021. 1

    Pith/arXiv arXiv 2021

  5. [5]

    The magnitude vector of images.Journal of Applied and Computational Topology, 8(3):447–473, 2024

    Michael F Adamer, Edward De Brouwer, Leslie O’Bray, and Bastian Rieck. The magnitude vector of images.Journal of Applied and Computational Topology, 8(3):447–473, 2024

    2024

  6. [6]

    Metric space magnitude for evaluating the diversity of latent representations

    Katharina Limbeck, Rayna Andreeva, Rik Sarkar, and Bastian Rieck. Metric space magnitude for evaluating the diversity of latent representations. InAdvances in Neural Information Processing Systems. 2024

    2024

  7. [7]

    Geometry-aware edge pooling for graph neural networks

    Katharina Limbeck, Lydia Mezrag, Guy Wolf, and Bastian Rieck. Geometry-aware edge pooling for graph neural networks. InAdvances in Neural Information Processing Systems. 2026. 1

    2026

  8. [8]

    Comparative analysis of indicators for multi-objective diversity opti- mization

    Ksenia Pereverdieva, André Deutz, Tessa Ezendam, Thomas Bäck, Hèrm Hofmeyer, and Michael Emmerich. Comparative analysis of indicators for multi-objective diversity opti- mization. InInternational Conference on Evolutionary Multi-Criterion Optimization, 2025. 1

    2025

  9. [9]

    The magnitude of dominated sets: a Pareto compliant indicator grounded in metric geometry.arXiv preprint arXiv:2604.18147, 2026

    Michael Emmerich. The magnitude of dominated sets: a Pareto compliant indicator grounded in metric geometry.arXiv preprint arXiv:2604.18147, 2026. 1

    Pith/arXiv arXiv 2026

  10. [10]

    Persistent magnitude for the quantitative analysis of the structure and stability of carboranes.Journal of Computational Biophysics and Chemistry, 23(06):741–751, 2024

    Wanying Bi, Xin Fu, Jingyan Li, and Jie Wu. Persistent magnitude for the quantitative analysis of the structure and stability of carboranes.Journal of Computational Biophysics and Chemistry, 23(06):741–751, 2024. 1

    2024

  11. [11]

    Topological magnitude for protein flexibility analysis.Royal Society Open Science, 12(12), 2025

    Wanying Bi, Hongsong Feng, Jie Wu, Jingyan Li, and Guo-Wei Wei. Topological magnitude for protein flexibility analysis.Royal Society Open Science, 12(12), 2025. 1

    2025

  12. [12]

    Categorifying the magnitude of a graph.Homology, Homotopy and Applications, 19(2):31–60, 2017

    Richard Hepworth and Simon Willerton. Categorifying the magnitude of a graph.Homology, Homotopy and Applications, 19(2):31–60, 2017. 1

    2017

  13. [13]

    Magnitude homology of enriched categories and metric spaces.Algebraic & Geometric Topology, 21(5):2175–2221, 2021

    Tom Leinster and Michael Shulman. Magnitude homology of enriched categories and metric spaces.Algebraic & Geometric Topology, 21(5):2175–2221, 2021. 1, 4

    2021

  14. [14]

    Blurred magnitude homology of functional connectome for ASD diagnosis.Frontiers in Psychiatry, 16: 1677282, 2025

    Alexander Kachura, Vsevolod Chernyshev, Oleg Kachan, and Egor Levchenko. Blurred magnitude homology of functional connectome for ASD diagnosis.Frontiers in Psychiatry, 16: 1677282, 2025. 1

    2025

  15. [15]

    Cambridge, 2002

    Allen Hatcher.Algebraic Topology. Cambridge, 2002. 1, 3

    2002

  16. [16]

    Springer Nature, 2026

    Michael Robinson.Computational Homological Algebra. Springer Nature, 2026. 1

    2026

  17. [17]

    Richard P Stanley.Enumerative Combinatorics, volume 1. 1997. 2

    1997

  18. [18]

    Poset topology: tools and applications

    Michelle Wachs. Poset topology: tools and applications. In Ezra Miller, Victor Reiner, and Bernd Sturmfels, editors,Geometric Combinatorics. AMS, 2007. 2, 6

    2007

  19. [19]

    Magnitude homology of metric spaces and order complexes.Bulletin of the London Mathematical Society, 53(3):893–905, 2021

    Ryuki Kaneta and Masahiko Yoshinaga. Magnitude homology of metric spaces and order complexes.Bulletin of the London Mathematical Society, 53(3):893–905, 2021. 2

    2021

  20. [20]

    Iterated magnitude homology.Advances in Mathematics, 468:110210, 2025

    Emily Roff. Iterated magnitude homology.Advances in Mathematics, 468:110210, 2025. 2

    2025

  21. [21]

    Fast Möbius and zeta transforms

    Tommaso Pegolotti, Bastian Seifert, and Markus Püschel. Fast Möbius and zeta transforms. arXiv preprint arXiv:2211.13706, 2022. 2

    Pith/arXiv arXiv 2022

  22. [22]

    Magnitude cohomology.Mathematische Zeitschrift, 301(4):3617–3640,

    Richard Hepworth. Magnitude cohomology.Mathematische Zeitschrift, 301(4):3617–3640,

  23. [23]

    Geometric approach to graph magnitude homology

    Yasuhiko Asao and Kengo Izumihara. Geometric approach to graph magnitude homology. Homology, Homotopy and Applications, 23(1):297–310, 2020. 2

    2020

  24. [24]

    Magnitude meets persistence: homology theories for filtered simplicial sets

    Nina Otter. Magnitude meets persistence: homology theories for filtered simplicial sets. Homology, Homotopy and Applications, 24(2):365–387, 2022. 2

    2022

  25. [25]

    Quantales, persistence, and magnitude homology.arXiv preprint arXiv:1910.02905, 2019

    Simon Cho. Quantales, persistence, and magnitude homology.arXiv preprint arXiv:1910.02905, 2019

    Pith/arXiv arXiv 1910

  26. [26]

    Persistent magnitude.Journal of Pure and Applied Algebra, 225(3):106517, 2021

    Dejan Govc and Richard Hepworth. Persistent magnitude.Journal of Pure and Applied Algebra, 225(3):106517, 2021. 2, 5

    2021

  27. [27]

    Magnitude homology and path homology.Bulletin of the London Mathematical Society, 55(1):375–398, 2023

    Yasuhiko Asao. Magnitude homology and path homology.Bulletin of the London Mathematical Society, 55(1):375–398, 2023. 2 10 Magnitude homology and Euler characteristics of DAGs

    2023

  28. [28]

    Path homologies of deep feedforward networks

    Samir Chowdhury, Thomas Gebhart, Steve Huntsman, and Matvey Yutin. Path homologies of deep feedforward networks. InIEEE International Conference On Machine Learning And Applications (ICMLA), 2019. 4, 6

    2019

  29. [29]

    A unified framework for numerically inverting Laplace trans- forms.INFORMS Journal on Computing, 18(4):408–421, 2006

    Joseph Abate and Ward Whitt. A unified framework for numerically inverting Laplace trans- forms.INFORMS Journal on Computing, 18(4):408–421, 2006. 5

    2006

  30. [30]

    An improved Talbot method for numerical Laplace transform inversion.Numerical Algorithms, 68(1):167–183, 2015

    Benedict Dingfelder and JAC Weideman. An improved Talbot method for numerical Laplace transform inversion.Numerical Algorithms, 68(1):167–183, 2015. 5

    2015

  31. [31]

    Springer, 2018

    Gerlind Plonka, Daniel Potts, Gabriele Steidl, and Manfred Tasche.Numerical Fourier Analysis. Springer, 2018. 5

    2018

  32. [32]

    Optimally scaled and optimally conditioned Vandermonde and Vandermonde- like matrices.BIT Numerical Mathematics, 51(1):103–125, 2011

    Walter Gautschi. Optimally scaled and optimally conditioned Vandermonde and Vandermonde- like matrices.BIT Numerical Mathematics, 51(1):103–125, 2011. 5

    2011

  33. [33]

    Johns Hopkins, 2013

    Gene H Golub and Charles F Van Loan.Matrix Computations. Johns Hopkins, 2013. 5

    2013

  34. [34]

    On Lovász’ lattice reduction and the nearest lattice point problem.Combinatorica, 6(1):1–13, 1986

    László Babai. On Lovász’ lattice reduction and the nearest lattice point problem.Combinatorica, 6(1):1–13, 1986. 5

    1986

  35. [35]

    Effects of the LLL reduction on the success probability of the Babai point and on the complexity of sphere decoding.IEEE Transactions on Information Theory, 59(8):4915–4926, 2013

    Xiao-Wen Chang, Jinming Wen, and Xiaohu Xie. Effects of the LLL reduction on the success probability of the Babai point and on the complexity of sphere decoding.IEEE Transactions on Information Theory, 59(8):4915–4926, 2013. 5

    2013

  36. [36]

    Springer, 2026

    Rubén Ballester, Carles Casacuberta, and Sergio Escalera.Topological Data Analysis for Neural Networks. Springer, 2026. 6

    2026

  37. [37]

    Bridging the black box: a survey on mechanistic interpretability in AI.ACM Computing Surveys, 58(8):1–35, 2026

    Shriyank Somvanshi, Md Monzurul Islam, Amir Rafe, Anannya Ghosh Tusti, Arka Chakraborty, Anika Baitullah, Tausif Islam Chowdhury, Nawaf Alnawmasi, Anandi Dutta, and Subasish Das. Bridging the black box: a survey on mechanistic interpretability in AI.ACM Computing Surveys, 58(8):1–35, 2026. 8

    2026

  38. [38]

    Euler calculus with applications to signals and sensing.Advances in Applied and Computational Topology, 70:75–145, 2012

    Justin Curry, Robert Ghrist, and Michael Robinson. Euler calculus with applications to signals and sensing.Advances in Applied and Computational Topology, 70:75–145, 2012. 9

    2012

  39. [39]

    Justin Curry, Sayan Mukherjee, and Katharine Turner. How many directions determine a shape and other sufficiency results for two topological transforms.Transactions of the American Mathematical Society, Series B, 9(32):1006–1043, 2022

    2022

  40. [40]

    An invitation to the Euler characteristic transform.The American Mathemati- cal Monthly, 132(1):15–25, 2025

    Elizabeth Munch. An invitation to the Euler characteristic transform.The American Mathemati- cal Monthly, 132(1):15–25, 2025

    2025

  41. [41]

    Topology meets machine mearning: an introduction using the Euler characteristic transform.Notices of the American Mathematical Society, 72:719–727, 2025

    Bastian Rieck. Topology meets machine mearning: an introduction using the Euler characteristic transform.Notices of the American Mathematical Society, 72:719–727, 2025. 9

    2025

  42. [42]

    not unique source at 1

    Steve Huntsman. Peel neighborhoods.arXiv preprint arXiv:2603.26645, 2026. 9 A Fast Euler characteristic code A.1 Floating-point code function chi = magnitude_homology_euler_st(adj) % Magnitude homology Euler characteristic of a st-DAG with adjacency matrix % adj, unique source 1, and unique target size(adj,1). % % Uses a reasonable Ansatz to mitigate the ...

    Pith/arXiv arXiv 2026