Pith. sign in

REVIEW 4 major objections 5 minor 40 references

Lossy compression of weighted graph adjacency matrices by transform coding

T0 review · 4 major / 5 minor · reviewed 2026-08-02 · deepseek-v4-flash

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

Pith's one-line read Weighted graphs compress best when edge weights are treated as a signal on the line graph, the paper argues.

desk verdict Solid extension of line-graph transform coding with a correct new identity, but the smoothness-difficulty claim is overreaching and Table 1's baseline is confounded. read the letter →

arxiv 2607.14834 v1 pith:2POKP3P7 submitted 2026-07-16 cs.IT eess.SPmath.IT

Kenta Yanagiya , Junya Hara , Hiroshi Higashi , Yuichi Tanaka , Antonio Ortega This is my paper
classification cs.ITeess.SPmath.IT
keywords weightedgraphcompressionlossylinefilterbanktransformcodingedgesmoothnesssignalprocessingadjacencymatrix
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that lossy compression of a weighted graph's edge weights should be done by first converting the graph to its line graph, so that each edge becomes a node and the edge weights become a graph signal. The authors propose a transform-coding pipeline: a critically-sampled graph filter bank on the line graph, scalar quantization, and entropy coding, while the binary topology is sent losslessly. They claim this achieves higher reconstruction SNR at equal bits per edge than direct preprocessing of the adjacency matrix on synthetic and real-world graphs, including power grids with non-geometric weights. They also introduce a normalized edge-weight variation, computable in O(N^2) without forming the line graph, and provide evidence that it predicts how many bits are needed. If correct, this gives a principled, topology-preserving way to trade fidelity against storage in weighted networks.

What carries the argument

The line graph G_e (nodes = edges of the original graph, edges = pairs of original edges sharing a node) and the graph filter bank applied on it. The line graph turns the edge-weight vector w into a graph signal; the filter bank (any critically-sampled graph filter bank can be plugged in) produces subband coefficients that are quantized and entropy-coded. The second mechanism is the edge-smoothness functional Delta_Le(w) = w^T L_e w, which Proposition 1 rewrites as a sum over original nodes of degree-squared times local weight variance, computable in O(N^2) without explicitly building the line graph, and used as a predictor of compression difficulty.

What would settle it

Take a fixed graph topology, assign edge weights independently from a wide distribution (so line-graph smoothness is minimal), and compare the proposed filter-bank pipeline against Direct-W: if the SNR-vs-bits-per-edge curves coincide, the line-graph smoothness premise fails. Equivalently, finding a real-world graph whose measured normalized edge variation is at or above the random baseline and yet still yields no BPE reduction would refute the claimed relationship.

Watch

Extended reading notes

Core claim

The central claim is that the line graph provides a canonical, permutation-invariant organization of edge weights, and that real-world edge weights are smooth on it, so a graph filter bank yields sparse coefficients. Concretely, after splitting the adjacency matrix into a binary topology A and edge-weight vector w, w is lifted to a signal on the line graph G_e and transformed by a critically-sampled filter bank; subband coefficients are uniformly quantized and Huffman-coded. Experiments with distance-based sensor and traffic graphs, an Erdős–Rényi model with controlled smoothness, and power-grid data with voltage edge weights show the proposed methods outperform Direct-GFB, Direct-W, and Ord

Load-bearing premise

The method's edge is conditional on typical edge weights being locally smooth — edges sharing a node tend to have similar weights — so that the line-graph transform concentrates energy; if weights are independent or anti-correlated, the transform offers no compression gain over direct coding.

Editorial extensions

If this is right

  • Weighted graphs with locally smooth edge weights can be stored or transmitted at a given fidelity with fewer bits per edge than direct quantization, while keeping the binary topology intact.
  • Because the topology is preserved losslessly, downstream graph-signal tasks such as denoising degrade less under compression than with methods that drop edges.
  • Any critically-sampled graph filter bank can be substituted into the framework, so improvements in graph transform design directly improve this compression scheme.
  • The edge-weight variation measure gives a cheap pre-compression estimate of how hard a given weighted graph is to compress, independent of the chosen filter bank.
  • The advantage holds for both geometric (distance-based) and non-geometric (voltage) edge weights, widening the applicability beyond sensor and traffic networks.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The line-graph smoothness premise suggests a natural adaptive scheme the paper does not explore: allocate more bits to local neighborhoods with high weight variance and fewer to smooth ones, potentially improving rate–distortion performance further.
  • The same lift-to-line-graph idea could be applied to other edge attributes or extended to directed graphs by replacing the line graph with a directed version, where the adjacency structure changes but the transform-coding principle would remain.
  • Since Proposition 1 computes smoothness from node-degree weighted local variance, one could in principle set the quantization step per graph based only on those local statistics, without ever constructing the line graph, enabling large-scale applications.
  • A direct ceiling test would compare against a transform learned from data on the line graph; if a learned basis beats fixed filter banks by a wide margin, the limiting factor is the filter-bank choice rather than the line-graph lift itself.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes a lossy compression framework for weighted undirected graphs in which the binary topology is sent losslessly and the edge-weight vector is compressed by mapping it to a graph signal on the line graph of the unweighted topology, applying a critically-sampled graph filter bank, scalar quantization, and entropy coding. The authors also introduce a normalized edge-weight variation Δ_L_e(w) (Eq. (10)) as a predicted measure of compression difficulty, with a local decomposition in Proposition 1 (Eq. (12)) that avoids explicit line-graph construction. Experiments on synthetic sensor, 10-NN, ER, traffic, and power-grid graphs compare the proposed method with direct matrix quantization, direct graph filter bank, and ordered DCT baselines, reporting higher reconstruction SNR at equal bits per edge and improved downstream denoising performance. The central technical derivation (Proposition 1) is correct, but several claims about smoothness validation, topology preservation, and the comparison protocol need attention.

Significance. If the claims hold, the line-graph transform-coding approach provides a principled, permutation-invariant organization of edge weights for lossy graph compression, and the local edge-variation formula is a useful, cheaply computable predictor of compressibility. The use of multiple filter banks and a downstream denoising task is a strength, and Proposition 1 is a clean contribution. The paper does not provide machine-checked proofs or code, but the mathematics in Section 4 is straightforward and verifiable. The main reservations concern the empirical baseline for smoothness, the consistency of the 'lossless topology' claim, and the fairness of the Direct-GFB rate comparison. These are fixable within the manuscript's scope.

major comments (4)
  1. [§4.2, Table 1] The smoothness evidence is confounded. The baseline uses independent uniform[0,1] weights, while physical weights (e.g., exp(-d²/0.3) in Eq. (11)) have a different marginal distribution with much smaller variance. The normalized variation Δ_L_e(w)/||w||² is not invariant to marginal scale/variance, so lower values for physical weights do not establish smoothness on the line graph. The correct null is a random permutation of the same multiset of physical weights over the same topology. Please replace or supplement the baseline and re-evaluate the 'many realistic networks exhibit smooth variation' claim.
  2. [§3.1 and §5.2.2] The framework claims the binary topology is transmitted losslessly, but §3.1 states that if a reconstructed edge weight is zero, the edge is removed from the reconstructed adjacency matrix. Section 5.2.2 then explicitly attributes DSNR degradation to coarse quantization removing edges. Thus topology is not strictly preserved in reconstruction. This is a conceptual inconsistency in a central claim; specify how zero-valued reconstructed weights are handled (e.g., a small positive floor, a separate edge mask, or a revised claim) and adjust the abstract/introduction accordingly.
  3. [§4.4] The assertion that Δ_L_e(w) 'serves as a measure of compression difficulty' is supported by a standard transform-coding heuristic and qualitative observations (e.g., Fig. 6), not by a derivation or a quantitative relationship to the rate-distortion tradeoff. As σ increases, both the marginal variance and the line-graph variation increase, so the experiments do not isolate the role of Δ_L_e(w). To justify the stated 'theoretical basis' in the introduction, provide a formal link or a quantitative correlation (e.g., BPE at fixed SNR vs. Δ_L_e across many graph/weight pairs), or weaken the claim to a heuristic predictor.
  4. [§5.1.1, Direct-GFB] Direct-GFB is described as a baseline that does not separately transmit the binary topology, and its rate excludes BA. However, the graph filter bank used in Eq. (14) is defined on the graph Laplacian of the unweighted topology; the decoder cannot invert the transform without knowing that topology. If the topology is assumed as shared side information, then the proposed method's BA should be excluded or counted consistently for all methods. As written, the bitrate comparison for Direct-GFB may be unfair or the method may be undecodable. Please clarify the decoder side-information model and adjust the rate accounting.
minor comments (5)
  1. [Eq. (10)] The same symbol Δ_L_e(w) is used for both the unnormalized variation in Eq. (9) and the normalized variation in Eq. (10). Use a distinct notation (e.g., an overline or superscript) to avoid ambiguity.
  2. [Table 1] The row λ_e,max is not defined or discussed in the text. Clarify what it represents and why it is included.
  3. [§4.3] The claimed O(N²) cost for computing Eq. (12) is imprecise. Computing the per-node sums is O(|E|) once node degrees are known; the expression O(N²) only follows from the worst-case bound |E|=O(N²). State the actual complexity in terms of |E| or degrees.
  4. [Eq. (12)] The local statistics [µ_w]_i are undefined for isolated nodes with degree 0. Add a convention (e.g., the corresponding term is zero) or restrict to graphs with no isolated nodes.
  5. [Figures 9 and 10] The captions refer to 'black solid and dotted lines,' while Figure 5 uses 'solid and dashed.' Use consistent terminology to identify the losslessly compressed weighted and binary adjacency matrix rate lines.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the line-graph transform-coding pipeline is an independent construction; the Table 1 baseline concern is a validity issue, not a circular reduction.

full rationale

The paper's core claim is a construction, not a derivation that feeds its own output back in. Edge weights are mapped to a graph signal on the line graph (Eqs. (3)-(5)), transformed by independently published graph filter banks (Eq. (7)), quantized, entropy coded, and inverted (Eq. (8)); the reported SNR/BPE comparisons are against external baselines (Direct-W, Direct-GFB, Ordered-DCT, Binary). No parameter is fitted to the target data, and no prediction is defined in terms of the proposed smoothness measure. Delta_Le(w) in Eq. (9) is a direct statistic of w and the line graph, and Proposition 1 is an algebraic identity rather than a fitted relation; its use as a compression-difficulty predictor is motivated by transform-coding theory and tested in Section 5, not asserted by definition of the bitrate. The manuscript's self-references ([1] and the filter-bank papers [26,28,29,33], and [39] for an elementary degree formula) point to published, externally archived work and are not invoked as uniqueness theorems or as the sole evidence for the central claim. The most serious weakness is external to circularity: Section 4.2, Table 1 compares normalized variation against uniform[0,1] random weights, which does not control for the marginal distribution of physical weights (e.g., exp(-d^2/0.3) values are concentrated near 1), so the empirical claim that real-world edge weights are smooth on the line graph is not cleanly established. That is a validity/confounding concern about the empirical premise, not a reduction of the paper's result to its inputs.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The paper introduces no fitted parameters or new physical entities. The method's load-bearing assumptions are the empirical smoothness of edge weights on line graphs and the applicability of existing graph filter banks to line graphs; the smoothness-rate-distortion link is a heuristic, not a theorem. One experimental modeling constant, the kernel width 0.3, affects the smoothness evidence.

free parameters (1)
  • Gaussian kernel width for distance-based edge weights (Eq. 11) = 0.3
    Used to generate synthetic sensor, 10-NN, and traffic edge weights; the smoothness measurements in Table 1 and the rate-distortion curves depend on this hand-chosen scale. It is not fitted to compression results, but it is an adjustable modeling constant.
assumptions (5)
  • standard math For a simple undirected graph, the line graph degree satisfies [D_e]_alpha = d_i + d_j - 2, giving the local-global decomposition in Proposition 1.
    This identity, cited to [38,39], underpins Eq. (12). It holds only for simple graphs without self-loops or multi-edges.
  • domain assumption Signals with small quadratic variation on a graph have energy concentrated in low-frequency subspaces, so transform coding yields sparse coefficients.
    Relies on classical transform coding intuition; for arbitrary graph filter banks and quantization, rate-distortion optimality is not guaranteed.
  • domain assumption The real-world edge weights tested (distance-based traffic, voltage power-grid) are representative of typical weighted graphs; smoothness observed in Table 1 generalizes.
    The claim that edge smoothness serves as a practical measure of compression difficulty rests on this empirical premise.
  • domain assumption The graph filter banks from [26,28,29,33] can be applied to the line graph and achieve perfect reconstruction under critical sampling, including bipartition via Harary decomposition.
    The framework assumes these existing filter banks work on line graphs; for non-bipartite graphs the bipartition step may require special handling not discussed.
  • domain assumption Huffman coding after scalar quantization is a rate-distortion competitive entropy code, and the reported BPE calculation (Sec. 5.1.1) accounts for side information appropriately.
    The fairness of the bitrate comparison depends on this coding choice; the paper describes but does not fully formalize the overhead accounting.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lossy compression of weighted graph adjacency matrices by transform coding." pith.science (2026). https://pith.science/paper/2POKP3P7

@misc{pith2026260714834,
  author       = {Pith},
  title        = {Pith review of: Lossy compression of weighted graph adjacency matrices by transform coding},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2POKP3P7}},
  note         = {Machine review of arXiv:2607.14834}
}
read the original abstract

In this paper, we propose a compression framework for weighted graphs in which the graph topology is transmitted losslessly and edge weights are compressed lossily. A challenge in the lossy compression of edge weights is that the underlying relationships between edges are ambiguous. To address this issue, we first transform the unweighted graph into the corresponding line graph, whose nodes represent the edges of the original graph and whose edges encode the relationships between them. The line graph transform allows us to regard edge weights as a graph signal defined on the line graph. Instead of transmitting the edge-weight vector, we first transform it with a graph filter bank on the line graph. Then, quantization and entropy coding are performed on the transformed coefficients of the edge weight vector. In addition to the lossy compression method, we formalize edge smoothness on the line graph and show that it serves as a measure of the difficulty of compression. The proposed smoothness measure can be easily calculated without converting to a line graph. This provides insight into the expected compression performance of a given weighted graph. Experiments on synthetic and real-world data validate the effectiveness of the proposed method by comparing it with existing matrix preprocessing methods.

Figures

Figures reproduced from arXiv: 2607.14834 by the authors.

Figure 1
Figure 1. Illustration of line graph conversion. The red cir [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. Overall framework of the proposed weighted adjace [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Illustration of edge weights with different smooth [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Examples of synthetic graphs with N = 500. 5.1 Synthetic Graphs 5.1.1 Setup In the experiments on synthetic graphs, we use the following weighted graphs with N = 500: random sensor graph with |E| = 1, 781, 10- nearest-neighbor (10-NN) graph of Swiss Roll point cloud wi…
Figure 5
Figure 5. Figure 5: Performance comparison of SNRs of the reconstruct [PITH_FULL_IMAGE:figures/full_fig_p017_5.png]
Figure 6
Figure 6. Figure 6: Performance comparison of SNRs of the reconstruct [PITH_FULL_IMAGE:figures/full_fig_p018_6.png]
Figure 7
Figure 7. Figure 7: Traffic networks used in the real-world experiment. [PITH_FULL_IMAGE:figures/full_fig_p020_7.png]
Figure 8
Figure 8. Figure 8: Chilean power-grids used in the real-world experi [PITH_FULL_IMAGE:figures/full_fig_p021_8.png]
Figure 9
Figure 9. Figure 9: Performance comparison of SNRs of the reconstruct [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]
Figure 10
Figure 10. Figure 10: Performance comparison of SNRs of the reconstruc [PITH_FULL_IMAGE:figures/full_fig_p023_10.png]
Figure 11
Figure 11. Figure 11: The performance comparison of DSNRs of the graph s [PITH_FULL_IMAGE:figures/full_fig_p024_11.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

40 extracted references · 6 canonical work pages

  1. [1]

    Lossy Compression of Adjacency Matrices by Graph Filter Ba nks

    K. Yanagiya, J. Hara, H. Higashi, Y. Tanaka, and A. Ortega , “Lossy Compression of Adjacency Matrices by Graph Filter Ba nks”, in IEEE Int. Conf. Acoust. Speech Signal Process. 2024, 9386–90, doi: 10.1109/ICASSP48485.2024.10448045

    arXiv 2024

  2. [2]

    Graph Signal Processing: Overview, Challeng es, and Applications

    A. Ortega, P. Frossard, J. Kovačević, J. M. F. Moura, and P . Van- dergheynst, “Graph Signal Processing: Overview, Challeng es, and Applications”, Proc. IEEE, 106(5), 2018, 808–28, doi: 10.1109/JPROC.2018.2820

    arXiv 2018

  3. [3]

    The Emerging Field of Signal Processing on Graphs: Extending High-Dimensional Data Analysis to Netwo rks and Other Irregular Domains

    D. I. Shuman, S. K. Narang, P. Frossard, A. Ortega, and P. Vandergheynst, “The Emerging Field of Signal Processing on Graphs: Extending High-Dimensional Data Analysis to Netwo rks and Other Irregular Domains”, IEEE Signal Process. Mag., 30(3), 2013, 83–98, doi: 10.1109/MSP.2012.2235192

    arXiv 2013

  4. [4]

    Spectral Domain Sampling of Graph Signals

    Y. Tanaka, “Spectral Domain Sampling of Graph Signals”, IEEE Trans. Signal Process., 66(14), 2018, 3752–67, doi: 10.1109/TSP.2018.2839620

    arXiv 2018

  5. [5]

    Samplin g Signals on Graphs: From Theory to Applications

    Y. Tanaka, Y. C. Eldar, A. Ortega, and G. Cheung, “Samplin g Signals on Graphs: From Theory to Applications”, IEEE Signal Process. Mag., 37(6), 2020, 14–30, doi: 10.1109/MSP.2020.3016908. 26 Yanagiya et al

    arXiv 2020

  6. [6]

    Time-Varying Graph Learning with Constraints on Graph Temporal Variation

    K. Yamada, Y. Tanaka, and A. Ortega, “Time-Varying Graph Learning with Constraints on Graph Temporal Variation”, arXiv preprint arXiv:2001.03346, 2020

    arXiv 2001

  7. [7]

    Eigen- decomposition Free Sampling Set Selection for Graph Signal s

    A. Sakiyama, Y. Tanaka, T. Tanaka, and A. Ortega, “Eigen- decomposition Free Sampling Set Selection for Graph Signal s”, IEEE Trans. Signal Process., 67(10), 2019, 2679–92, doi: 10.1109/TSP.2019.29

    work page doi:10.1109/tsp.2019.29 2019

  8. [8]

    Graph Sig- nal Denoising via Trilateral Filter on Graph Spectral Domai n

    M. Onuki, S. Ono, M. Yamagishi, and Y. Tanaka, “Graph Sig- nal Denoising via Trilateral Filter on Graph Spectral Domai n”, IEEE Trans. Signal Inf. Process. Netw. , 2(2), 2016, 137–48, doi: 10.1109/TSIPN.2016.2532464

    arXiv 2016

Show all 40 references
  1. [9]

    Survey and Taxonomy of Lossless Graph Compression and Space-Efficient Graph Representation s

    M. Besta and T. Hoefler, “Survey and Taxonomy of Lossless Graph Compression and Space-Efficient Graph Representation s”, arXiv preprint arXiv:1806.01799 , 2019

    2019 arXiv

  2. [10]

    A Com- pact Representation of Graph Databases

    S. Álvarez, N. R. Brisaboa, S. Ladra, and Ó. Pedreira, “A Com- pact Representation of Graph Databases”, in Proc. 8th Workshop Min. Learn. Graphs, 2010, 18–25, doi: 10.1145/1830252.1830255

    2010

  3. [11]

    ZipG: A Memory-Efficient Graph Store for Interactive Queries

    A. Khandelwal, Z. Yang, E. Ye, R. Agarwal, and I. Stoica, “ZipG: A Memory-Efficient Graph Store for Interactive Queries”, in Proc. Int. Conf. Manag. Data , 2017, 1149–64, doi: 10.1145/3035918.3064012

    2017

  4. [12]

    Efficient Graph Management Based on Bitmap Indices

    N. Martínez-Bazan, M. Á. Águila-Lorente, V. Muntés-Mu lero, D. Dominguez-Sal, S. Gómez-Villamor, and J.-L. Larriba-Pe y, “Efficient Graph Management Based on Bitmap Indices”, in Proc. 16th Int. Database Eng. Appl. Sysmposium , 2012, 110–9, doi: 10.1145/2351476.2351489

    2012

  5. [13]

    Log(Graph): A near-Optimal High-Performance Gra ph Representation

    M. Besta, D. Stanojevic, T. Zivic, J. Singh, M. Hoerold, and T. Hoefler, “Log(Graph): A near-Optimal High-Performance Gra ph Representation”, in Proc. 27th Int. Conf. Parallel Archit. Compil. Tech. 2018, 1–13, doi: 10.1145/3243176.3243198

    2018

  6. [14]

    Co mpres- sion of Weighted Graphs

    H. Toivonen, F. Zhou, A. Hartikainen, and A. Hinkka, “Co mpres- sion of Weighted Graphs”, in Proc. 17th ACM SIGKDD Int. Conf. Knowl. Discov. Data Min. 2011, 965–73, doi: 10.1145/2020408.2020566

    2011

  7. [15]

    Edge Sampling of Graphs Based on Edge Smoothness

    K. Yanagiya, K. Yamada, Y. Katsuhara, T. Takatani, and Y . Tanaka, “Edge Sampling of Graphs Based on Edge Smoothness”, in IEEE Int. Conf. Acoust. Speech Signal Process. 2022, 5932–6, doi: 10.1109/ICASSP43922.2022.9747724

    2022

  8. [16]

    Spectral Sparsification of Graphs

    D. A. Spielman and S.-H. Teng, “Spectral Sparsification of Graphs”, SIAM J. Comput., 40(4), 2011, 981–1025, doi: 10.1137/08074489X. Lossy compression of weighted graph adjacency matrices by t ransform coding 27

    2011 doi

  9. [17]

    Graph Coarsening with Pr e- served Spectral Properties

    Y. Jin, A. Loukas, and J. JaJa, “Graph Coarsening with Pr e- served Spectral Properties”, in Proc. 23rd Int. Conf. Artif. Intell. Stat. 2020, 4452–62

    2020

  10. [18]

    Accelerating Sparse Mat rix Com- putations via Data Compression

    J. Willcock and A. Lumsdaine, “Accelerating Sparse Mat rix Com- putations via Data Compression”, in Proc. 20th Annu. Int. Conf. Supercomput. 2006, 307, doi: 10.1145/1183401.1183444

    2006

  11. [19]

    Simplification of N etworks by Edge Pruning

    F. Zhou, S. Mahler, and H. Toivonen, “Simplification of N etworks by Edge Pruning”, in Bisociative Knowledge Discovery: An In- troduction to Concept, Algorithms, Tools, and Application s, ed. M. R. Berthold, 2012, 179–98, doi: 10.1007/978-3-642-31830-6_13

    2012 doi

  12. [20]

    On Compressing Weighted Time-Evolving Graphs

    W. Liu, A. Kan, J. Chan, J. Bailey, C. Leckie, J. Pei, and R . Kotagiri, “On Compressing Weighted Time-Evolving Graphs” , in Proc. 21st ACM Int. Conf. Inf. Knowl. Manag. 2012, 2319–22, doi: 10.1145/2396761.2398630

    2012

  13. [21]

    Graph Spec- tral Image Processing

    G. Cheung, E. Magli, Y. Tanaka, and M. K. Ng, “Graph Spec- tral Image Processing”, Proc. IEEE , 106(5), 2018, 907–30, doi: 10.1109/JPROC.2018.2799702

    2018

  14. [22]

    Graph Signal Sampli ng Under Stochastic Priors

    J. Hara, Y. Tanaka, and Y. C. Eldar, “Graph Signal Sampli ng Under Stochastic Priors”, IEEE Trans. Signal Process. , 71, 2023, 1421–34, doi: 10.1109/TSP.2023.3267990

    2023

  15. [23]

    Multi-Channel Sampling on Graph s and Its Relationship to Graph Filter Banks

    J. Hara and Y. Tanaka, “Multi-Channel Sampling on Graph s and Its Relationship to Graph Filter Banks”, IEEE Open J. Signal Process., 4, 2023, 148–56, doi: 10.1109/OJSP.2023.3249112

    2023

  16. [24]

    A Tutorial on Spectral Clustering

    U. Von Luxburg, “A Tutorial on Spectral Clustering”, Stat Com- put, 17(4), 2007, 395–416, doi: 10.1007/s11222-007-9033-z

    2007 doi

  17. [25]

    Caldarelli and M

    G. Caldarelli and M. Catanzaro, Networks: A Very Short Intro- duction, 1st ed., No. 335, Oxford Univ. Press, 2012

    2012

  18. [26]

    Compact Support Biorthogon al Wavelet Filterbanks for Arbitrary Undirected Graphs

    S. K. Narang and A. Ortega, “Compact Support Biorthogon al Wavelet Filterbanks for Arbitrary Undirected Graphs”, IEEE Trans. Signal Process., 61(19), 2013, 4673–85, doi: 10.1109/TSP.2013.2273197

    2013

  19. [27]

    Oversampled Graph Laplacia n Ma- trix for Graph Filter Banks

    A. Sakiyama and Y. Tanaka, “Oversampled Graph Laplacia n Ma- trix for Graph Filter Banks”, IEEE Trans. Signal Process., 62(24), 2014, 6425–37, doi: 10.1109/TSP.2014.2365761

    2014

  20. [28]

    Two - Channel Critically Sampled Graph Filter Banks With Spectra l Domain Sampling

    A. Sakiyama, K. Watanabe, Y. Tanaka, and A. Ortega, “Two - Channel Critically Sampled Graph Filter Banks With Spectra l Domain Sampling”, IEEE Trans. Signal Process. , 67(6), 2019, 1447–60, doi: 10.1109/TSP.2019.2892033. 28 Yanagiya et al

    2019

  21. [29]

    Two Chan nel Filter Banks on Arbitrary Graphs With Positive Semi Definite Variation Operators

    E. Pavez, B. Girault, A. Ortega, and P. A. Chou, “Two Chan nel Filter Banks on Arbitrary Graphs With Positive Semi Definite Variation Operators”, IEEE Trans. Signal Process., 71, 2023, 917– 32, doi: 10.1109/TSP.2023.3257983

    2023

  22. [30]

    On the Linegraph of a Directed Graph

    M. Aigner, “On the Linegraph of a Directed Graph”, Math Z , 102(1), 1967, 56–61, doi: 10.1007/BF01110285

    1967 doi

  23. [31]

    Line Graphs of Weighted Ne t- works for Overlapping Communities

    T. S. Evans and R. Lambiotte, “Line Graphs of Weighted Ne t- works for Overlapping Communities”, Eur. Phys. J. B , 77(2), 2010, 265–72, doi: 10.1140/epjb/e2010-00261-8 , arXiv: 0912.4389

    2010 arXiv

  24. [32]

    Di screte Signal Processing on Graphs: Sampling Theory

    S. Chen, R. Varma, A. Sandryhaila, and J. Kovačević, “Di screte Signal Processing on Graphs: Sampling Theory”, IEEE Trans. Signal Process., 63(24), 2015, 6510–23, doi: 10.1109/TSP.2015.2469645

    2015

  25. [33]

    Perfect Reconstruction Two -Channel Wavelet Filter Banks for Graph Structured Data

    S. K. Narang and A. Ortega, “Perfect Reconstruction Two -Channel Wavelet Filter Banks for Graph Structured Data”, IEEE Trans. Signal Process., 60(6), 2012, 2786–99, doi: 10.1109/TSP.2012.2188718

    2012

  26. [34]

    Spectral Grap h Wavelets and Filter Banks With Low Approximation Error

    A. Sakiyama, K. Watanabe, and Y. Tanaka, “Spectral Grap h Wavelets and Filter Banks With Low Approximation Error”, IEEE Trans. Signal Inf. Process. Netw., 2(3), 2016, 230–45, doi: 10.1109/TSIPN.2016.25

    2016 doi

  27. [35]

    Wav elets on Graphs via Spectral Graph Theory

    D. K. Hammond, P. Vandergheynst, and R. Gribonval, “Wav elets on Graphs via Spectral Graph Theory”, ArXiv09123848 Cs Math , 2009, arXiv: 0912.3848 [cs, math]

    2009 arXiv

  28. [36]

    GSP-T raffic Dataset: Graph signal processing dataset based on traffic sim u- lation

    R. Kumagai, H. Kojima, H. Higashi, and Y. Tanaka, “GSP-T raffic Dataset: Graph signal processing dataset based on traffic sim u- lation”, in Graph Signal Process. Workshop , 2024

    2024

  29. [37]

    In-Depth Data on the Network Structure and Hourly Activity of the Central Chilean Power Grid

    H. Kim, D. Olave-Rojas, E. Álvarez-Miranda, and S.-W. S on, “In-Depth Data on the Network Structure and Hourly Activity of the Central Chilean Power Grid”, Sci Data, 5(1), 2018, 180209, doi: 10.1038/sdata.2018.209

    2018 doi

  30. [38]

    De- gree Distribution and Assortativity in Line Graphs of Compl ex Networks

    X. Wang, S. Trajanovski, R. E. Kooij, and P. Van Mieghem, “De- gree Distribution and Assortativity in Line Graphs of Compl ex Networks”, Physica A: Statistical Mechanics and its Applications , 445, 2016, 343–56, doi: 10.1016/j.physa.2015.10.109

    2016 doi

  31. [39]

    Edge Sampling of Graphs: Graph Signal Processing A p- proach With Edge Smoothness

    K. Yanagiya, K. Yamada, Y. Katsuhara, T. Takatani, and Y . Tanaka, “Edge Sampling of Graphs: Graph Signal Processing A p- proach With Edge Smoothness”, IEEE Trans. Signal Inf. Process. Netw., 11, 2025, 1592–604, doi: 10.1109/TSIPN.2025.3639969

    2025

  32. [40]

    The Biparticity of a Gr aph

    F. Harary, D. Hsu, and Z. Miller, “The Biparticity of a Gr aph”, J. Graph Theory, 1(2), 1977, 131–3, doi: 10.1002/jgt.3190010208

    1977 doi

Pith tools

Reviewed August 2, 2026 · model on record in the stance chip above.