Pith. sign in

REVIEW 4 major objections 4 minor 90 references

From communities to interpretable network and word embedding: an unified approach

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A vertex's embedding can be the share of its weighted degree that lands in each community, giving sparse, interpretable, and stable graph and word embeddings that predict degree and PageRank almost perfectly.

desk verdict The interpretability story is real, but the perfect degree-regression scores are mathematically impossible under Eq. 4.1, so the quantitative core needs a code audit before the headline claims can be trusted. read the letter →

arxiv 2412.08187 v1 pith:ZQZRMKD2 submitted 2024-12-11 cs.CL cs.AIcs.LG

Thibault Prouteau , Nicolas Dugué , Simon Guillot This is my paper · ORCID
classification cs.CLcs.AIcs.LG
keywords graphembeddingwordinterpretabilitycommunitydetectionnoderecallco-occurrencenetworksparserepresentationmatrixfactorization
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

The paper argues that grouping a network into communities can do double duty as an embedding method: instead of learning dense vectors with a black-box model, represent each vertex by the fraction of its weighted connections that go to each community. This gives a sparse, high-dimensional vector whose dimensions have direct meaning, because each dimension names a community. The authors implement the idea in two ways, SINr-NR (a direct normalized degree share) and SINr-MF (a matrix factorization), and show that SINr-NR predicts vertex degree and PageRank almost perfectly, matches standard baselines on link prediction and word similarity/categorization, and remains stable across runs. The reason to care is that interpretability by design and low computational cost come without a large performance penalty on the tasks tested.

What carries the argument

The load-bearing object is the Node Recall vector, $NR_i(u)=d_{C_i}(u)/d(u)$, where $d_{C_i}(u)$ is the total weight of edges from vertex $u$ to vertices in community $C_i$ and $d(u)$ is $u$'s weighted degree. The $i$-th coordinate is the share of $u$'s connectivity that flows into community $i$, so each dimension is a named community rather than a latent feature. The Louvain algorithm, a fast modularity-based community-detection heuristic with a resolution parameter $\gamma$, supplies the communities and controls the dimension count; the $NR$ formula then turns the partition into cheap sparse vectors. SINr-MF replaces the direct formula with gradient descent that factors the adjacency matrix into a community-membership matrix, giving a second implementation of the same bipartite-projection framework.

What would settle it

Build a degree-preserving randomized version of a word co-occurrence network and run the paper's word-intrusion protocol on SINr-NR embeddings; if human annotators still detect intruders as often as on the real network, then the apparent interpretability comes from vector arithmetic rather than from the community structure the framework claims to expose.

Watch

Extended reading notes

Core claim

The central claim is that a vertex's role in a network is captured by its normalized distribution of weighted degree over the detected communities, formally $NR_i(u)=d_{C_i}(u)/d(u)$ for community $C_i$. Collecting these shares across communities yields the embedding vector, with each coordinate tied to a tangible community, making the space auditable by construction. The paper reports near-perfect $R^2$ for vertex-degree and PageRank regression on five graphs, and competitive accuracy on link prediction, community classification, and, for word co-occurrence networks, word similarity and concept categorization. When applied to text, the strongest dimensions align with coherent topics, and the nearest-neighbor structure varies far less across repeated runs than that of dense or sparse interpretable baselines.

Load-bearing premise

The detected communities must be semantically coherent and stable, because each embedding dimension is only as meaningful as the community it represents.

Editorial extensions

If this is right

  • Vertex degree can be read off almost exactly from the spread of the NR vector, since the coordinates sum to one and their diffusion across communities encodes total degree.
  • PageRank becomes linearly predictable from the NR vector on the tested networks, indicating that community-degree distribution carries global influence information.
  • Word embeddings built from co-occurrence networks inherit the same interpretability, with polysemous words decomposing into separate topic dimensions.
  • The method's runtime is orders of magnitude below neural and random-walk baselines on the tested graphs, with SINr-NR the fastest.
  • The single resolution parameter $\gamma$ trades dimension count against downstream accuracy, giving a practical control knob.

Reading between the lines

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

  • If the community structure is coherent, the same dimensions could serve as an automatic topic labeler, for instance by representing a document through the NR-weighted communities that its words activate.
  • The near-perfect degree and PageRank regressions suggest NR vectors encode a centrality-like signature; one could test whether adding them as features improves link prediction or node-ranking beyond current baselines.
  • Because the only random element is community detection and the paper finds it stable, the framework should extend to diachronic text: embeddings from different time slices can be compared dimension by dimension once communities are aligned.
  • A sharper test of the mechanism would randomize edges while preserving degrees; if word-intrusion performance barely drops, interpretability would be a property of degree distributions rather than of community semantics.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper introduces Lower Dimension Bipartite Graph Framework (LDBGF), a general approach for interpretable graph and word embeddings based on projecting a graph onto a low-dimensional bipartite structure between vertices and detected communities. Two implementations are presented: SINr-NR, which uses the normalized community-degree distribution (node recall) of each vertex as its embedding, and SINr-MF, which factorizes the adjacency matrix against a community-membership matrix by gradient descent. The authors evaluate both on link prediction, vertex degree regression, clustering coefficient regression, community clustering/classification, PageRank regression, and on word similarity, concept categorization, stability, and interpretability tasks. The central claims are that SINr-NR produces high-quality, interpretable, and stable embeddings at very low computational cost, and that SINr-MF is competitive on classical graph tasks.

Significance. If the experimental results are correct, the paper makes a useful contribution: a fast, intrinsically interpretable embedding method that works for both graphs and word co-occurrence networks, with encouraging stability properties (Section 5.3.2 reports high NMI and low nearest-neighbor variation for SINr-NR). The word intrusion evaluation and the probing examples for 'insulin' and 'mint' also provide concrete evidence for the interpretability claim. However, the degree and PageRank regression results (Tables 3 and 7) are load-bearing for the 'high-quality graph embedding' claim, and the reported R^2 values appear mathematically impossible under the paper's own definition of the NR embedding. The paper therefore needs a careful audit and correction of these experiments before its central claims can be accepted.

major comments (4)
  1. [Section 5.2.1, Table 3 and Eq. (4.1)] The reported R^2 = 1.000 for vertex degree regression on Cora, Email-EU, arXiv, and Facebook is incompatible with the definition of SINr-NR. Equation (4.1) defines NR_i(u) = d_{C_i}(u) / d(u), so every embedding row lies on the simplex: the features contain only the relative distribution of a vertex's weighted degree across communities, not its scale d(u). Two vertices with identical community-degree proportions but different total degrees would have identical feature vectors, so a linear regression from these features cannot predict degree exactly. The explanation given in the text ('the more diffuse, the higher the potential degree') refers to spread, not scale, and cannot account for R^2 = 1.0. The result is the signature of an evaluation artifact, such as inadvertently using the unnormalized community degrees d_{C_i}(u) as features (in which case degree is simply the row sum) or of target leakage. Because this table is a principal piece of evidence for the 'high-quality graph embedding' claim, the authors should audit their implementation, report the corrected regression results, and explain the discrepancy.
  2. [Section 5.2.3, Table 7] The near-perfect PageRank regression results (R^2 between 0.949 and 0.991) are likely affected by the same issue as the degree regression. PageRank scores depend on the scale of a vertex's connections, and the normalized NR features discard that scale. Even if PageRank is strongly correlated with degree, a perfect or near-perfect linear fit from simplex-valued features is not credible. The authors should rerun this experiment with the same corrected features and report the actual R^2 values, or clarify why PageRank should be exactly recoverable from the normalized distribution.
  3. [Section 5.2.1, hyperparameter selection] The gamma parameter is selected using link prediction accuracy on the same graph (Figure 5) and then reused for all other experiments on that graph. This is a form of hyperparameter optimization on the evaluation data, and it is particularly problematic because the chosen gamma values are then used to claim superiority in degree and PageRank regression. Moreover, the text states that for Email-EU the maximum link prediction accuracy for SINr-NR is reached at gamma = 3, but the authors then say 'the gamma value chosen for Cora, Citeseer and Eu is 1'. This inconsistency needs to be resolved, and the authors should either use a separate validation set for gamma selection or clearly report the sensitivity of the other tasks to gamma.
  4. [Section 4.3 and Section 5 (SINr-MF parameters)] The paper states in several places that 'the only parameter of our methods is the gamma multi-resolution parameter' (e.g., Section 5.2.1). This is false for SINr-MF, which is trained with 3000 epochs, a learning rate of 5e-3, and stochastic gradient descent, as stated in Section 5. The claim should be corrected to reflect that SINr-NR has one main parameter (gamma) while SINr-MF has additional optimization hyperparameters.
minor comments (4)
  1. [Contributions list (Section 1)] The contribution 'SINr-NR runs in linear time' should be qualified: the method uses the Louvain community detection algorithm, which is quasi-linear in the number of edges but not strictly linear in all implementations. Please state the precise algorithmic complexity with respect to n and m.
  2. [Section 5.3.1, Table 8 and surrounding text] The text claims that 'SINr-NR results are indeed better than those of SPINE on MEN and WS353', which is true on both corpora for WS353 and on OANC for MEN, but on SCWS SPINE is better on OANC (0.44 vs 0.39) and the two are tied on BNC (0.54 vs 0.54). Please qualify the claim to avoid overstating the comparison.
  3. [Section 2.2 and Section 4.3] Louvain is described as 'unsupervised without parameters' in Section 4.3, but the paper later uses the multi-resolution parameter gamma from the same algorithm. Please make the description consistent, e.g., 'unsupervised in the sense of not requiring labeled data, but with a resolution parameter gamma'.
  4. [Tables 3 and 7] For completeness, the regression setup should specify whether an intercept term is used and whether features are standardized. This is important for interpreting R^2 values, especially when comparing across methods with different feature scales.

Circularity Check

2 steps flagged · score 6.0 of 10

Degree regression R²=1.000 is not explainable by Eq. 4.1's normalized NR vectors; it indicates the evaluation used unnormalized community degrees, making the 'prediction' an algebraic identity, and γ is tuned on the link-prediction test set before being reused elsewhere.

  1. self definitional [Section 4.3, Eq. (4.1); Section 5.2.1, 'Vertex degree regression']
    "NRi(u) = dCi(u)/d(u) with dCi(u) =∑ v∈Ci Ωuv (4.1) ... we proceed to predict the degree of vertices in the graph ... the target value for each vertex is extracted from the degree sequence of G."

    Under Eq. 4.1, each embedding row is the normalized community-degree distribution, so the target d(u) is exactly the sum of the unnormalized coordinates d_Ci(u). A linear regression from the unnormalized community degrees would give R²=1.000 because the target is then the row sum of the features. With the normalized vectors of Eq. 4.1, R²=1.000 is impossible: two vertices with the same community proportions but different total degrees share identical features and must receive identical predictions. Table 3's R²=1.000 on four of five graphs is therefore the signature that the implementation used the unnormalized community degrees as input, i.e., the degree 'prediction' reduces, by construction, to summing the features.

  2. fitted input called prediction [Section 5.2.1, Figure 5 and following paragraph]
    "As a consequence of these results, the γ value chosen for Cora, Citeseer and Eu is 1, when γ is set to 5 for arXiv and Facebook for all the experiments on these networks."

    The only hyperparameter γ is selected by maximizing accuracy on the link-prediction task of the same graphs (Figure 5 plots 'the accuracy on the link prediction task according to the γ value'), and this same γ is then fixed for all other experiments on those networks, including degree and PageRank regression. Thus the link-prediction numbers in Table 2 are test-set tuned rather than independent predictions, and the downstream experiments inherit a γ chosen on a related evaluation target on the same datasets. This does not by itself prove the degree regression is circular, but it removes the independence of the reported link-prediction performance and inflates the apparent advantage of the method.

full rationale

The most serious issue is the degree regression. Eq. 4.1 defines each embedding coordinate as d_{C_i}(u)/d(u), so the feature vector is a normalized version of the very community-degree decomposition whose sum is the target d(u). If the code uses the unnormalized d_{C_i}(u) as features, then the reported R²=1.000 is an algebraic identity, not a learned prediction; if it uses the published normalized vectors, R²=1.000 is mathematically impossible. Either way, the central quantitative claim 'SINr-NR can produce high-quality graph embeddings' rests on a result that reduces to the construction of the embedding. The γ-selection procedure compounds this by tuning the only hyperparameter on the link-prediction test metric and then reusing it globally, making the link-prediction results optimistically biased. I do not find load-bearing circularity in the word-embedding evaluations: those are compared against external similarity/categorization benchmarks and human intrusion judgments, and the cited prior work of the authors is used as a framework reference rather than as an imported uniqueness theorem. The PageRank regression is not directly circular because PageRank is not simply the row sum of the community-degree features, but its near-perfect R² is suspect in light of the degree leakage and should be audited with the released code.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central claim relies on the assumption that Louvain communities at a chosen gamma provide a stable, semantically coherent basis for representation. This is a domain assumption, not derived from first principles. The only fitted hyperparameter is gamma, which is tuned on the evaluation tasks. No new physical or mathematical entities are introduced.

free parameters (2)
  • gamma (Louvain resolution parameter) = 1 for Cora, Citeseer, Email-Eu; 5 for arXiv and Facebook; other values in the link prediction sweep
    Gamma controls the number of communities and thus the embedding dimension. It is selected per network based on link prediction accuracy (Section 5.2.1, Figure 5), and then reused in all other experiments on that network, making it a tuned hyperparameter rather than a fixed constant.
  • SINr-MF learning rate and epochs = learning rate 5e-3, 3000 epochs
    These are chosen hyperparameters for gradient descent (Section 5). They are not fitted to data but are arbitrary choices that affect convergence and performance.
assumptions (4)
  • domain assumption Communities detected by Louvain on a graph correspond to semantically meaningful groups in both social and word co-occurrence networks.
    The entire interpretability claim rests on this. The paper states communities are 'clusters of related words that co-occur frequently together and are thus semantically intertwined' (Section 5.3), but this is assumed rather than proven for each dataset.
  • domain assumption The distributional hypothesis: words with similar contexts have similar meanings.
    Used to justify word embeddings from co-occurrence networks (Sections 1 and 5.3). This is a standard assumption in NLP.
  • domain assumption The PMI threshold (A(u,v)=0 when PMI<0) preserves the semantically relevant edges in word co-occurrence graphs.
    Equation 5.3 removes all negative PMI co-occurrences. The paper claims this 'helps to better detect communities', but the choice of threshold 0 is not derived and may discard useful signal.
  • standard math The minimum edge clique cover problem is NP-hard, so communities are a valid relaxation.
    Section 4.1 states the edge clique cover is NP-hard (citing Erdos et al. 1988) and therefore uses communities as a relaxation. This is a standard computational complexity argument.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From communities to interpretable network and word embedding: an unified approach." pith.science (2026). https://pith.science/paper/ZQZRMKD2

@misc{pith2026241208187,
  author       = {Pith},
  title        = {Pith review of: From communities to interpretable network and word embedding: an unified approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZQZRMKD2}},
  note         = {Machine review of arXiv:2412.08187}
}
read the original abstract

Modelling information from complex systems such as humans social interaction or words co-occurrences in our languages can help to understand how these systems are organized and function. Such systems can be modelled by networks, and network theory provides a useful set of methods to analyze them. Among these methods, graph embedding is a powerful tool to summarize the interactions and topology of a network in a vectorized feature space. When used in input of machine learning algorithms, embedding vectors help with common graph problems such as link prediction, graph matching, etc. Word embedding has the goal of representing the sense of words, extracting it from large text corpora. Despite differences in the structure of information in input of embedding algorithms, many graph embedding approaches are adapted and inspired from methods in NLP. Limits of these methods are observed in both domains. Most of these methods require long and resource greedy training. Another downside to most methods is that they are black-box, from which understanding how the information is structured is rather complex. Interpretability of a model allows understanding how the vector space is structured without the need for external information, and thus can be audited more easily. With both these limitations in mind, we propose a novel framework to efficiently embed network vertices in an interpretable vector space. Our Lower Dimension Bipartite Framework (LDBGF) leverages the bipartite projection of a network using cliques to reduce dimensionality. Along with LDBGF, we introduce two implementations of this framework that rely on communities instead of cliques: SINr-NR and SINr-MF. We show that SINr-MF can perform well on classical graphs and SINr-NR can produce high-quality graph and word embeddings that are interpretable and stable across runs.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

90 extracted references · 70 canonical work pages

  1. [1]

    & Poesio, M

    Almuhareb, A. & Poesio, M. (2005) Concept learning and categorization from the web. In proceedings of the annual meeting of the Cognitive Science society, volume 27

    2005

  2. [2]

    & Kruszewski, G

    Baroni, M., Dinu, G. & Kruszewski, G. (2014) Don’t count, predict! A systematic comparison of context- counting vs. context-predicting semantic vectors. In ACL, pages 238–247

    2014

  3. [3]

    & Lenci, A

    Baroni, M. & Lenci, A. (2011) How we BLESSed distributional semantic evaluation.GEometrical Models of Natural Language Semantics, pages 1–10. FROM COMMUNITIES TO INTERPRETABLE NETWORK AND WORD EMBEDDING 37 of 40

    2011

  4. [4]

    & Niyogi, P

    Belkin, M. & Niyogi, P. (2001) Laplacian eigenmaps and spectral techniques for embedding and clustering. Advances in neural information processing systems, 14

    2001

  5. [5]

    K., Meneni, K., Danisch, M., Guillaume, J.-L

    Bhowmick, A. K., Meneni, K., Danisch, M., Guillaume, J.-L. & Mitra, B. (2020) LouvainNE: Hierarchical Louvain Method for High Quality and Scalable Network Embedding. In WSDM, pages 43–51

    2020

  6. [6]

    D., Guillaume, J

    Blondel, V . D., Guillaume, J. L., Lambiotte, R. & Lefebvre, E. (2008) Fast unfolding of communities in large networks. Journal of Statistical Mechanics: Theory and Experiment. arXiv: 0803.0476

    arXiv 2008

  7. [7]

    & Rosvall, M

    Bohlin, L., Edler, D., Lancichinetti, A. & Rosvall, M. (2014) Community detection and visualization of networks with the map equation framework. Measuring scholarly impact: methods and practice, pages 3–34

    2014

  8. [8]

    & Mikolov, T

    Bojanowski, P., Grave, E., Joulin, A. & Mikolov, T. (2017) Enriching Word Vectors with Subword Informa- tion. Transactions of the Association for Computational Linguistics

    2017

Show all 90 references
  1. [9]

    & Page, L

    Brin, S. & Page, L. (1998) The anatomy of a large-scale hypertextual Web search engine.Computer Networks and ISDN Systems, 30(1-7), 107–117

    1998

  2. [10]

    & Velcin, J

    Brochier, R., Guille, A. & Velcin, J. (2019) Global Vectors for Node Representations. InThe World Wide Web Conference, pages 2587–2593. arXiv:1902.11004 [cs]

    2019 arXiv

  3. [11]

    Broniatowski, D. A. et al. (2021) Psychological foundations of explainability and interpretability in artificial intelligence. NIST, Tech. Rep

    2021

  4. [12]

    D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A

    Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A. et al. (2020) Language models are few-shot learners. Neurips, 33, 1877–1901

    2020

  5. [13]

    Bruni, E., Tran, N. K. & Baroni, M. (2014) Multimodal Distributional Semantics.Journal of Artificial Intelli- gence Research, 49, 1–47

    2014

  6. [14]

    Cao, S., Lu, W. & Xu, Q. (2015) GraRep: Learning Graph Representations with Global Structural Information. In CIKM, pages 891–900

    2015

  7. [15]

    Cao, S., Lu, W. & Xu, Q. (2016) Deep neural networks for learning graph representations. In Proceedings of the AAAI conference on artificial intelligence, volume 30

    2016

  8. [16]

    & Park, N

    Chakraborty, T., Cui, Z. & Park, N. (2018) Metadata vs. Ground-truth: A Myth behind the Evolution of Community Detection Methods. In WWW’18, pages 45–46, Lyon, France. ACM Press

    2018

  9. [17]

    & Blei, D

    Chang, J., Gerrish, S., Wang, C., Boyd-graber, J. & Blei, D. (2009) Reading Tea Leaves: How Humans Interpret Topic Models. In Neurips, volume 22

    2009

  10. [18]

    Chen, J., Za ¨ıane, O. R. & Goebel, R. (2008) An Unsupervised Approach to Cluster Web Search Results Based on Word Sense Communities. In International Conference on Web Intelligence and Intelligent Agent Technology, volume 1, pages 725–729

    2008

  11. [19]

    & Largeron, C

    Choudhary, M., Laclau, C. & Largeron, C. (2022) A survey on fairness for machine learning on graphs.arXiv preprint arXiv:2205.05396

    2022 arXiv

  12. [20]

    (2007) British National Corpus, XML edition

    Consortium, B. (2007) British National Corpus, XML edition. Oxford Text Archive

    2007

  13. [21]

    & Lenca, P

    Dao, V .-L., Bothorel, C. & Lenca, P. (2017) Community detection methods can discover better structural clusters than ground-truth communities. In ASoNAM’17, pages 395–400, Sydney Australia. ACM

    2017

  14. [22]

    & Toutanova, K

    Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. (2019) BERT: Pre-training of Deep Bidirectional Trans- formers for Language Understanding. In North American Chapter of the Association for Computational Lin- guistics, pages 4171–4186

    2019

  15. [23]

    & Perez, A

    Dugu´e, N., Lamirel, J.-C. & Perez, A. (2019) Bringing a feature selection metric from machine learning to complex networks. In Complex Networks and Their Applications VII, pages 107–118

    2019

  16. [24]

    T., Nguyen, Q

    Duong, C. T., Nguyen, Q. V . H. & Aberer, K. (2019) Interpretable node embeddings with mincut loss. In Learning and Reasoning with Graph-Structured Representations Workshop-ICML

    2019

  17. [25]

    & Ordman, E

    Erd¨oos, P., Faudree, R. & Ordman, E. T. (1988) Clique partitions and clique coverings.Discrete Mathematics, 72(1-3), 93–101

    1988

  18. [26]

    Data & Description : http://wordspace.collocations.de/doku.php/data:esslli2008:start

    ESSLLI (2008) Shared Tasks from the ESSLLI 2008 Workshop. Data & Description : http://wordspace.collocations.de/doku.php/data:esslli2008:start

    2008

  19. [27]

    & Smith, N

    Faruqui, M., Tsvetkov, Y ., Yogatama, D., Dyer, C. & Smith, N. A. (2015) Sparse Overcomplete Word Vector Representations. In ACL, pages 1491–1500

    2015

  20. [28]

    & Ruppin, E

    Finkelstein, L., Gabrilovich, E., Matias, Y ., Rivlin, E., Solan, Z., Wolfman, G. & Ruppin, E. (2001) Placing 38 of 40 T. PROUTEAU, N. DUGU´E, S. GUILLOT search in context: The concept revisited. WWW, pages 406–414

    2001

  21. [29]

    & Barth´elemy, M

    Fortunato, S. & Barth´elemy, M. (2007) Resolution limit in community detection.Proceedings of the National Academy of Sciences, 104(1), 36–41

    2007

  22. [30]

    & Zou, J

    Garg, N., Schiebinger, L., Jurafsky, D. & Zou, J. (2018) Word embeddings quantify 100 years of gender and ethnic stereotypes. Proceedings of the National Academy of Sciences, 115(16), E3635–E3644

    2018

  23. [31]

    A., McClure, D., Glorieux, F., Reboul, M., Porter, J

    Gefen, A., Algee-Hewitt, M. A., McClure, D., Glorieux, F., Reboul, M., Porter, J. & Riguet, M. (2017) Vector based measure of semantic shifts across different cultural corpora as a proxy to comparative history of ideas. JADH 2017, page 12

    2017

  24. [32]

    & Newman, M

    Girvan, M. & Newman, M. E. (2002) Community structure in social and biological networks. Proceedings of the national academy of sciences, 99(12), 7821–7826

    2002

  25. [33]

    & Leskovec, J

    Grover, A. & Leskovec, J. (2016) node2vec: Scalable Feature Learning for Networks. arXiv:1607.00653 [cs, stat]

    2016 arXiv

  26. [34]

    & Latapy, M

    Guillaume, J.-L. & Latapy, M. (2004) Bipartite structure of all complex networks. Information processing letters, 90(5), 215–221

    2004

  27. [35]

    L., Leskovec, J

    Hamilton, W. L., Leskovec, J. & Jurafsky, D. (2016) Cultural shift or linguistic drift? comparing two compu- tational measures of semantic change. In EMNLP, page 2116

    2016

  28. [36]

    Harris, Z. S. (1954) Distributional structure. Word, 10(2-3), 146–162. Publisher: Taylor & Francis

    1954

  29. [37]

    H., Socher, R., Manning, C

    Huang, E. H., Socher, R., Manning, C. D. & Ng, A. Y . (2012) Improving word representations via global context and multipleword prototypes. In ACL

    2012

  30. [38]

    & Johnson, K

    Kim, M., Kim, J. & Johnson, K. (2023) Race, Gender, and Age Biases in Biomedical Masked Language Models. In ACL, pages 11806–11815

    2023

  31. [39]

    (2013) Multi-scale modularity and dynamics in complex networks

    Lambiotte, R. (2013) Multi-scale modularity and dynamics in complex networks. In Dynamics On and Of Complex Networks, Volume 2: Applications to Time-Varying Dynamical Systems, pages 125–141

    2013

  32. [40]

    & Kert ´esz, J

    Lancichinetti, A., Fortunato, S. & Kert ´esz, J. (2009) Detecting the overlapping and hierarchical community structure in complex networks. New Journal of Physics, 11(3), 033015

    2009

  33. [41]

    & Fortunato, S

    Lancichinetti, A., Kivel ¨a, M., Saram ¨aki, J. & Fortunato, S. (2010) Characterizing the Community Structure of Complex Networks. PLOS ONE, 5(8), 1–8

    2010

  34. [42]

    Lancichinetti, A., Radicchi, F., Ramasco, J. J. & Fortunato, S. (2011) Finding Statistically Significant Com- munities in Networks. PLOS ONE, 6(4), 1–18

    2011

  35. [43]

    & Inouye, M

    Lannelongue, L., Grealey, J. & Inouye, M. (2021) Green Algorithms: Quantifying the Carbon Footprint of Computation. Advanced Science, 8(12), 2100707

    2021

  36. [44]

    H., Newman, D

    Lau, J. H., Newman, D. & Baldwin, T. (2014) Machine Reading Tea Leaves: Automatically Evaluating Topic Coherence and Topic Model Quality. In EACL, pages 530–539

    2014

  37. [45]

    & Cunningham, P

    Lee, C. & Cunningham, P. (2014) Community detection: effective evaluation on large social networks.Journal of Complex Networks, 2(1), 19–37

    2014

  38. [46]

    Lee, D. D. & Seung, H. S. (1999) Learning the parts of objects by nonnegative matrix factorization. Nature, 401, 788–791

    1999

  39. [47]

    & Goldberg, Y

    Levy, O. & Goldberg, Y . (2014) Neural Word Embedding as Implicit Matrix Factorization. In Neurips, vol- ume 27

    2014

  40. [48]

    & Dagan, I

    Levy, O., Goldberg, Y . & Dagan, I. (2015) Improving distributional similarity with lessons learned from word embeddings. ACL, 3, 211–225

    2015

  41. [49]

    Liu, F., Yang, X., Guan, N. & Yi, X. (2016) Online graph regularized non-negative matrix factorization for large-scale datasets. Neurocomputing, 204, 162–171

    2016

  42. [50]

    & Stoyanov, V

    Liu, Y ., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L. & Stoyanov, V . (2019) RoBERTa: A Robustly Optimized BERT Pretraining Approach. arXiv:1907.11692 [cs]

    2019 arXiv

  43. [51]

    Lundberg, S. M. & Lee, S.-I. (2017) A unified approach to interpreting model predictions.Advances in neural information processing systems, 30

    2017

  44. [52]

    & Subelj, L

    Makarov, I., Kiselev, D., Nikitinsky, N. & Subelj, L. (2021) Survey on graph embeddings and their applica- tions to machine learning problems on graphs. PeerJ Computer Science, 7, e357

    2021

  45. [53]

    & Dean, J

    Mikolov, T., Chen, K., Corrado, G. & Dean, J. (2013) Efficient estimation of word representations in vector FROM COMMUNITIES TO INTERPRETABLE NETWORK AND WORD EMBEDDING 39 of 40 space. In ICLR. arXiv: 1301.3781

    2013 arXiv

  46. [54]

    Miller, G. A. (1995) WordNet: A Lexical Database for English. Commun. ACM, 38(11), 39–41

    1995

  47. [55]

    Murphy, B., Talukdar, P. P. & Mitchell, T. (2012) Learning effective and interpretable semantic models using non-negative sparse embedding. In COLING, pages 1933–1950

    2012

  48. [56]

    Nancy Ide, Randi Reppen, K. S. (2011) The Open ANC (OANC). ORTOLANG

    2011

  49. [57]

    M., Tworek, J., Yuan, Q., Tezak, N., Kim, J

    Neelakantan, A., Xu, T., Puri, R., Radford, A., Han, J. M., Tworek, J., Yuan, Q., Tezak, N., Kim, J. W., Hallacy, C. et al. (2022) Text and code embeddings by contrastive pre-training.arXiv preprint arXiv:2201.10005

    2022 arXiv

  50. [58]

    & Fort, K

    N´ev´eol, A., Dupont, Y ., Bezanc ¸on, J. & Fort, K. (2022) French CrowS-Pairs: Extending a challenge dataset for measuring social bias in masked language models to a language other than English. InACL, pages 8521–8531

    2022

  51. [59]

    E., Suci, G

    Osgood, C. E., Suci, G. J. & Tannenbaum, P. H. (1957) The measurement of meaning. The measurement of meaning. Univer. Illinois Press, Oxford, England. Pages: 342

    1957

  52. [60]

    & Zhu, W

    Ou, M., Cui, P., Pei, J., Zhang, Z. & Zhu, W. (2016) Asymmetric Transitivity Preserving Graph Embedding. In SIGKDD, pages 1105–1114

    2016

  53. [61]

    Palmer, S. E. (1977) Hierarchical structure in perceptual representation.Cognitive Psychology, 9(4), 441–474

    1977

  54. [62]

    Panigrahi, A., Simhadri, H. V . & Bhattacharyya, C. (2019) Word2Sense: Sparse Interpretable Word Embed- dings. In ACL, pages 5692–5705

    2019

  55. [63]

    & Dean, J

    Patterson, D., Gonzalez, J., Le, Q., Liang, C., Munguia, L.-M., Rothchild, D., So, D., Texier, M. & Dean, J. (2021) Carbon emissions and large neural network training. arXiv preprint arXiv:2104.10350

    2021 arXiv

  56. [64]

    Peel, L., Larremore, D. B. & Clauset, A. (2017) The ground truth about metadata and community detection in networks. Science advances, 3(5), e1602548

    2017

  57. [65]

    & Manning, C

    Pennington, J., Socher, R. & Manning, C. D. (2014) Glove: Global vectors for word representation. In EMNLP, pages 1532–1543

    2014

  58. [66]

    & Skiena, S

    Perozzi, B., Al-Rfou, R. & Skiena, S. (2014) DeepWalk: Online Learning of Social Representations. SIGKDD, pages 701–710. arXiv: 1403.6652

    2014 arXiv

  59. [67]

    & Skiena, S

    Perozzi, B., Kulkarni, V ., Chen, H. & Skiena, S. (2017) Don’t Walk, Skip! Online Learning of Multi-scale Network Embeddings. In ASONAM, pages 258–265

    2017

  60. [68]

    (2020) Qualitative Evaluation of Word Embeddings: Investigating the Instability in Neural- Based Models

    Pierrejean, B. (2020) Qualitative Evaluation of Word Embeddings: Investigating the Instability in Neural- Based Models. PhD thesis, Universit´e Toulouse 2 - Jean Jaur`es

    2020

  61. [69]

    & Meignier, S

    Prouteau, T., Connes, V ., Dugu´e, N., Perez, A., Lamirel, J.-C., Camelin, N. & Meignier, S. (2021) SINr: Fast Computing of Sparse Interpretable Node Representations is not a Sin!. In IDA, pages 325–337

    2021

  62. [70]

    & Meignier, S

    Prouteau, T., Dugu´e, N., Camelin, N. & Meignier, S. (2022) Are Embedding Spaces Interpretable? Results of an Intrusion Detection Evaluation on a Large French Corpus. In LREC 2022, page 4414–4419

    2022

  63. [71]

    & Liu, P

    Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y ., Li, W. & Liu, P. J. (2020) Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer. arXiv:1910.10683

    2020 arXiv

  64. [72]

    N., Albert, R

    Raghavan, U. N., Albert, R. & Kumara, S. (2007) Near linear time algorithm to detect community structures in large-scale networks. Physical Review E - Statistical, Nonlinear, and Soft Matter Physics. arXiv: 0709.2938

    2007 arXiv

  65. [73]

    T., Singh, S

    Ribeiro, M. T., Singh, S. & Guestrin, C. (2016) ”Why Should I Trust You?”: Explaining the Predictions of Any Classifier. arXiv:1602.04938 [cs, stat]

    2016 arXiv

  66. [74]

    Roweis, S. T. & Saul, L. K. (2000) Nonlinear dimensionality reduction by locally linear embedding. science, 290(5500), 2323–2326. Publisher: American Association for the Advancement of Science

    2000

  67. [75]

    & Sutton, C

    Rozemberczki, B., Davies, R., Sarkar, R. & Sutton, C. (2020) GEMSEC: graph embedding with self cluster- ing

    2020

  68. [76]

    & Goodenough, J

    Rubenstein, H. & Goodenough, J. B. (1965) Contextual correlates of synonymy.Communications of the ACM, 8(10), 627–633

    1965

  69. [77]

    (2019) Stop explaining black box machine learning models for high stakes decisions and use inter- pretable models instead

    Rudin, C. (2019) Stop explaining black box machine learning models for high stakes decisions and use inter- pretable models instead. Nature machine intelligence, 1(5), 206–215

    2019

  70. [78]

    & Yao, X

    Serra, G., Xu, Z., Niepert, M., Lawrence, C., Tino, P. & Yao, X. (2021) Interpreting Node Embedding with Text-labeled Graphs. In IJCNN, pages 1–8

    2021

  71. [79]

    & Malik, J

    Shi, J. & Malik, J. (2000) Normalized Cuts and Image Segmentation. IEEE Transactions on pattern analysis and machine intelligence, 22(8). 40 of 40 T. PROUTEAU, N. DUGU´E, S. GUILLOT

    2000

  72. [80]

    & Vaudaine, R

    Sinha, A., Cazabet, R. & Vaudaine, R. (2019) Systematic biases in link prediction: comparing heuristic and graph embedding based methods. In Complex Networks, pages 81–93

    2019

  73. [81]

    & Meyerhenke, H

    Staudt, C., Sazonovs, A. & Meyerhenke, H. (2014) NetworKit: An Interactive Tool Suite for High- Performance Network Analysis. CoRR, abs/1403.3005

    2014 arXiv

  74. [82]

    & McCallum, A

    Strubell, E., Ganesh, A. & McCallum, A. (2020) Energy and policy considerations for modern deep learning research. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 13693–13696

    2020

  75. [83]

    & Hovy, E

    Subramanian, A., Pruthi, D., Jhamtani, H., Berg-Kirkpatrick, T. & Hovy, E. (2018) SPINE: SParse Inter- pretable Neural Embeddings. In Thirty-Second AAAI Conference on Artificial Intelligence

    2018

  76. [84]

    & Mei, Q

    Tang, J., Qu, M., Wang, M., Zhang, M., Yan, J. & Mei, Q. (2015) LINE: Large-scale Information Network Embedding. In WWW, pages 1067–1077

    2015

  77. [85]

    B., Silva, V

    Tenenbaum, J. B., Silva, V . d. & Langford, J. C. (2000) A Global Geometric Framework for Nonlinear Dimen- sionality Reduction. Science, 290(5500), 2319–2323

    2000

  78. [86]

    & Liu, T.-Y

    Tian, F., Dai, H., Bian, J., Gao, B., Zhang, R., Chen, E. & Liu, T.-Y . (2014) A probabilistic model for learning multi-prototype word embeddings. In COLING, pages 151–160

    2014

  79. [87]

    & M¨uller, E

    Tsitsulin, A., Mottin, D., Karras, P. & M¨uller, E. (2018) VERSE: Versatile Graph Embeddings from Similarity Measures. In WWW, pages 539–548. arXiv:1803.04742 [cs]

    2018 arXiv

  80. [88]

    & Zhu, W

    Wang, D., Cui, P. & Zhu, W. (2016) Structural Deep Network Embedding. InSIGKDD, pages 1225–1234

    2016

  81. [89]

    & Yang, S

    Wang, X., Cui, P., Wang, J., Pei, J., Zhu, W. & Yang, S. (2017) Community Preserving Network Embedding. Proceedings of the AAAI Conference on Artificial Intelligence, 31(1)

    2017

  82. [90]

    & Chang, E

    Yang, C., Liu, Z., Zhao, D., Sun, M. & Chang, E. Y . (2015) Network representation learning with rich text information.. In IJCAI, volume 2015, pages 2111–2117

    2015

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.