RAE: A Rule-Driven Approach for Attribute Embedding in Property Graph Recommendation

Junwei Hu; Michael Bewong; Selasi Kwashie; Sibo Zhao; Zaiwen Feng

REVIEW 5 major objections 5 minor 37 references

RAE: A Rule-Driven Approach for Attribute Embedding in Property Graph Recommendation

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Rule-guided random walks on property graphs lift top-20 recommendation by 10.6%.

desk verdict The rule-guided random walk idea is genuinely new and the ablation supports it, but the paper's evaluation is untestable as written because the datasets are never converted into a defined recommendation task. read the letter →

arxiv 2506.08314 v2 pith:JLXYE4PK submitted 2025-06-10 cs.IR

Sibo Zhao , Michael Bewong , Selasi Kwashie , Junwei Hu , Zaiwen Feng This is my paper

classification cs.IR

keywords recommendationsystemspropertygraphsgraphassociationrulesattributeembeddingrandomwalkwithrestartconvolutionalnetworksruleminingtop-n

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 recommendation systems underuse attribute information when attributes are treated as simple triples . It proposes RAE, which first mines graph association rules from a property graph—dependencies such as x.a -> y.b—then uses those rules to guide random walks over an attribute-extended graph, yielding node and attribute embeddings that encode both structure and semantics. These embeddings are fed into a LightGCN-style recommender. The paper reports that RAE outperforms five baselines on four datasets, with an average improvement of 10.6% in both Recall@20 and NDCG@20, and that it degrades less than baselines when attributes are randomly removed. If true, RAE shows a practical path to extracting more value from attribute-rich graphs in recommender systems.

What carries the argument

The load-bearing mechanism is the rule-based random walk on the attribute-extended graph G', which augments the original property graph with attribute nodes and association edges. A restart probability alpha controls walk length; whenever the walk arrives at an attribute, mined graph association rules are checked, and if a rule's LHS is satisfied the walk jumps to the RHS, adding node-attribute pairs beyond those reachable by ordinary walks. These pairs define forward affinity F and backward affinity B, both shifted PMI scores, which are then factored into low-dimensional embeddings Xf, Xb, Y by minimizing a squared reconstruction objective. The final node embedding is a weighted, normalized combination of forward and backward embedding products with the attribute embedding Y, and this combination is fed into a LightGCN-style GCN as its input layer.

What would settle it

Run RAE with the mined rules replaced by random attribute-to-attribute jumps (same number of jumps, same walk lengths, but rule targets shuffled): if Recall@20 and NDCG@20 stay at the same level as the full RAE, then the specific rule content is not what drives the gains and only the extra jump mechanism matters. Similarly, if a rule miner with high support and confidence thresholds yields no rules, the method should collapse to RAEn; observing otherwise would show the improvement comes from embedding refinement rather than from the rules.

Watch

Extended reading notes

Core claim

The central claim is that rule-guided random walks on property graphs produce attribute embeddings that improve top-20 recommendation beyond what triple-based attribute integration or plain attributed-network random walks achieve. Specifically, RAE mines rules of the form Q[x](X -> Y), where the pattern Q[x] describes a subgraph and X, Y are conjunctions of attribute and equality literals; one example is 'users who share a major, and one likes a genre, support recommending the movie'. These rules are applied during a random walk with restart on an attribute-extended graph G': when the walk lands on an attribute satisfying a rule's left-hand side, it jumps to the rule's right-hand side, generating extra node-attribute pairs. Forward and backward affinities are estimated as shifted pointwise mutual information from the resulting pair sets, and embeddings are learned by minimizing squared reconstruction error under a dimension budget. The embeddings are combined and passed to a LightGCN-style model trained with a BPR objective. The paper reports state-of-the-art Recall@20 and NDCG@20 on Facebook, Blogcatalog, Flickr, and Citeseer, with the largest gains on Citeseer (29.64% in Recall@20), and shows smaller performance drops when up to 80% of attributes are removed.

Load-bearing premise

The load-bearing premise is that the rule-mining module produces reliable, meaningful graph association rules; the paper delegates the mining to a GAR-style mechanism without specifying the inference algorithm, support and confidence thresholds, or how left-hand-side propagation works, so if the mined rules are noisy or trivial the rule jumps add little beyond ordinary random walks.

Editorial extensions

If this is right

If RAE is correct, recommender systems can improve top-20 recall and ranking by mining rules from the dataset's own property graph instead of relying on external knowledge graphs.
Rule-guided attribute embeddings are more robust to missing attributes than triple-based attribute integration: on Citeseer, Recall@20 barely drops from 0.161 to 0.160 when 80% of attributes are removed.
The improvement holds across sparsity levels, meaning attributes can compensate for scarce interaction data in cold and sparse user groups.
RAE's design makes recommendations traceable to mined semantic rules, which supports explainability without sacrificing accuracy.

Reading between the lines

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

Beyond the paper: the reported 10.6% average improvement is dominated by the Citeseer result, where rule guidance likely matters most because the graph is directed and sparse; on denser social graphs the gain over the best baseline is roughly 3-6%, so the practical benefit may vary by graph type.
Beyond the paper: because the rule mining module is delegated to an unspecified GAR-style miner, the method's reproducibility depends on the mining implementation; a natural test is to vary support and confidence thresholds and measure how rule count and quality affect the jump contribution.
Beyond the paper: the random-walk embedding could be plugged into contrastive or knowledge-graph-based recommenders, since the rule-guided embeddings are orthogonal to contrastive augmentation and external KG construction.
Beyond the paper: a direct comparison against RippleNet or RuleRec on a graph with an external KG would clarify whether internal property-graph rules can substitute for curated external knowledge.

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.

Desk Editor's Note

read the letter

The core idea here is worth taking seriously. RAE mines graph association rules from a property graph and uses them as jump instructions in a random walk, producing attribute embeddings that are then fed into a LightGCN-style recommender. That rule-guided walk is a real extension of PANE's plain random walks, and the ablation (RAE vs RAEn) suggests the rules contribute beyond the walks themselves. The paper is also honest about building on GAR and PANE rather than claiming those pieces as new. For what it is—an incremental but genuine integration—the writing is clear and the method section is mostly readable.

The problem is that the empirical claim is not currently supported. The stress-test note lands hard: Table 1 lists Facebook, BlogCatalog, Flickr, and Citeseer as if they were recommendation datasets, but the paper never defines how these homogeneous attributed graphs become a user-item interaction set O+. The evaluation uses Recall@20 and NDCG@20 with BPR training, yet there is no description of which nodes are users, which are items, or how positive and negative pairs are sampled. Without that, every number in Tables 2 and 3 is floating. The abstract's "10.6% average improvement in both R@20 and NDCG@20" is also contradicted by Table 2's own arithmetic: the average R@20 improvement is 11.31%, and the average NDCG@20 improvement is 7.30%. That mismatch suggests carelessness rather than malice, but it still needs fixing.

The second soft spot, which the reader flagged, is the rule-mining black box. Section 4.1 explicitly declines to specify the inference mechanism or thresholds, yet the entire novelty over PANE depends on those rules being meaningful. If the rules are noisy or trivial, the gains in Table 3 shrink to noise. This is a genuine load-bearing gap, though it is secondary to the missing task definition.

Minor points: no error bars or variance reporting; hyperparameters are listed but no sensitivity analysis; several graph-dependency references are from the same group (Refs 10, 11, 28, 31), but that does not drive the central claim, so I would not call it a citation-pattern problem.

Who is this for? Researchers working on attribute-aware graph recommendation or property-graph embedding. The idea could inspire a useful variant. But in its current form it is not reproducible, and the headline result is untestable. A serious referee should get it, but the path forward is a major revision: define the interaction construction explicitly, report error bars, reconcile the abstract with the table, and specify the rule-mining setup. I would not cite it until those are addressed.

Referee Report

5 major / 5 minor

Summary. The paper proposes RAE, a three-module framework for recommendation on property graphs: (1) a rule-mining module that extracts graph association rules (GAR) from attribute literals, variable literals, and constant literals; (2) a rule-based random walk embedding module that extends PANE-style random walks with rule jumps to compute forward and backward shifted-PMI affinities, then factorizes these affinities into vertex and attribute embeddings; and (3) a recommendation module that feeds normalized forward/backward embeddings into a LightGCN-style model trained with a BPR pairwise loss. The authors evaluate RAE on Facebook, BlogCatalog, Flickr, and Citeseer against BPR-MF, LightGCN, IMP-GCN, AF-GCN, and A2-GCN using Recall@20 and NDCG@20, and report gains over the best baseline in Table 2, robustness to sparsity and attribute missingness in Figure 3, and an ablation study in Table 3. The abstract claims an average improvement of 10.6% in both Recall@20 and NDCG@20.

Significance. If the claimed results are valid, RAE would make a useful contribution by showing that mined attribute-association rules can guide random-walk embeddings and improve top-20 recommendation over attribute-aware GCN baselines, while also offering a degree of interpretability through the mined rules. The paper builds on established components (PANE-style random walks, GAR rule mining, LightGCN), which is a reasonable design, and it releases a GitHub repository with code. The ablation study attempts to isolate the contribution of rule guidance versus plain random walks. However, the central claim is currently not testable from the manuscript because the evaluation protocol does not define how the four homogeneous attributed graphs are converted into a user-item recommendation task, the rule-mining component is described as a black box, and the headline improvement does not match the arithmetic of the reported table. These issues are load-bearing for the paper's core claim, so the manuscript needs substantial revision before its conclusions can be assessed.

major comments (5)

[Section 5.1, Table 1, and Definition 1] The evaluation protocol is undefined: the four datasets in Table 1 are homogeneous attributed networks (social/citation graphs) with no reported user set, item set, interaction set O+, or procedure for deriving the bipartite recommendation task. The text in Section 5.1 defines sparsity as the ratio of observed interactions to total possible user-item interactions, but Table 1 only reports |V|, |E_V|, |R|, and |E_R|, so the reported sparsity values cannot be checked. Since BPR-MF, LightGCN, and the GCN module are evaluated on Recall@20 and NDCG@20, the paper must specify exactly which edges or node pairs are treated as user-item interactions (e.g., whether all social/citation edges are reinterpreted as interactions, or whether some separate interaction matrix is used), and report |U|, |I|, |O+|, and the train/test split at the interaction level. Without this, Tables 2 and 3 and the 10.6% average-improvement claim are untestable.
[Abstract and Table 2] The claimed average improvement of 10.6% in both Recall@20 and NDCG@20 is not supported by the numbers in Table 2. Averaging the 'Improvement' row across the four datasets gives approximately 11.31% for Recall@20 and 7.30% for NDCG@20, and averaging all eight values gives approximately 9.30%. The authors should either correct the abstract, state precisely which baseline and which aggregation formula produce 10.6%, or remove the claim.
[Section 4.1 and Section 4.2] The rule-mining module is a black box and this is the main novelty over PANE. Section 4.1 says only that RAE 'leverages Graph Association Rules (GAR) [12]' and focuses on three literal types, explicitly declining to detail 'the full inference mechanisms of GAR'. Section 4.2 then relies on undefined operations: when a walk reaches an attribute, it checks whether a rule is satisfied, and if the LHS has multiple literals it 'first propagates within the LHS before jumping to the RHS'. The paper never defines the rule satisfaction test, the LHS-propagation procedure, the support/confidence thresholds, the number of mined rules per dataset, or the run-time of the mining step. Because the comparison of RAE versus RAEn in Table 3 is intended to demonstrate the value of rule guidance, the rule-mining and rule-jump semantics must be specified precisely enough for reproduction.
[Equation (5), Section 4.3] The transformation in Eq. (5) is notationally underspecified and likely dimensionally inconsistent. The forward embedding X_f[vi] and backward embedding X_b[vi] are vectors in R^{k/2}, while Y[rj] is also in R^{k/2}, but the equation writes an elementwise-style product X_f ⊙ Y and claims X_final ∈ R^{n×d}, where d is the number of attributes. It is unclear how per-node vectors of dimension k/2 become an n×d matrix, how the attribute index in Y is selected for each node, and what operation '⊙' denotes in this context. The authors should define all symbols and give the exact shapes of X_f, X_b, Y, and X_final, or the implementation is not reproducible from the paper.
[Section 5.2, Table 2] The numerical results are reported without variance, confidence intervals, or significance testing. The per-dataset gains over the best baseline are mostly 3–6%, except for the 29.64% Recall@20 gain on Citeseer. Given the small margins on BlogCatalog and Flickr, the absence of repeated-run statistics or paired significance tests makes it difficult to determine whether the improvement over AF-GCN and A2-GCN is meaningful rather than due to random initialization or hyperparameter tuning. The authors should report means and standard deviations over multiple runs, or otherwise provide a statistical basis for the comparison.

minor comments (5)

[Figure 3 caption] There is a typo in the caption: 'Perfomance' should be 'Performance'.
[Table 2, 'Improvement' row] The row is labeled 'Improvement' but the baseline is not explicitly identified; the text says the best baseline is AF-GCN, yet in several columns (e.g., Facebook R@20 and NDCG@20, Citeseer R@20) the best baseline is A2-GCN or LightGCN. Please state for each column which baseline the improvement is computed against.
[Section 5.1] The hyperparameter search is reported only as ranges ('explored learning rates from 0.01, 0.001, 0.0001'; L2 in [10^-5, 10^-2]). Please report the selected values for each dataset, since the final performance may be sensitive to these choices.
[Section 5.4, Figure 3(c)-(d)] The robustness claim for attribute missingness is supported only by qualitative descriptions in the text; the figure shows curves, but the exact values at each missing ratio are not tabulated, and the claim that 'RAE's Recall@20 drops slightly from 0.161 to 0.160 at 80%' should be verifiable in the figure. Please include the numeric values or a table.
[Section 4.2, Eq. (2)] The justification of Eq. (2) as a shifted PMI score contains an apparent mismatch: the denominator term under the logarithm uses P(r_j) with a factor n/(n+δ), but the text writes the formula directly in terms of p'_f. Please make the derivation explicit, including how δ enters the PMI expression, so that the formula can be reproduced.

Circularity Check

0 steps flagged · score 2.0 of 10

No load-bearing circularity: the RAE derivation is an unsupervised embedding pipeline (rule-guided RWR affinities, low-rank factorization, GCN/BPR) with the target labels entering only at the final BPR loss; self-citations appear in related work but are not load-bearing.

full rationale

I walked the chain from Section 4.1 to Eq. (6). Section 4.2 defines forward/backward affinities F and B (Eqs. 2-3) as shifted PMI scores over rule-guided random walks; Eq. (4) factorizes these affinities into Xf, Xb, Y; Eq. (5) normalizes the combination; Eq. (6) trains the GCN with BPR on observed interactions. The recommendation labels enter only in Eq. (6), so no fitted parameter is renamed as a prediction. The rule miner is external: Section 4.1 says 'Our approach leverages Graph Association Rules (GAR) [12]' and 'rather than detailing the full inference mechanisms of GAR'; this is an underspecified black-box dependency and a reproducibility gap, not a self-citation or a construction that forces the result. The self-cited works (Refs 10, 11, 28, 31) appear only in the related-work sentence 'similar to methods like graph association rules (GARs) [8,12], graph functional dependencies (GFDs) [9], graph differential dependencies (GDDs) [10,31], and graph entity dependencies (GEDs) [11,28]'; they supply no equation, uniqueness theorem, or forced choice in RAE, so they are not load-bearing. Two non-circular concerns are flagged for completeness: Section 5.1 does not specify how the homogeneous attributed networks (Facebook, BlogCatalog, Flickr, Citeseer) are converted into a user-item interaction set O+, and the abstract's '10.6%' average does not match the arithmetic of Table 2 (Recall improvements average 11.31%, NDCG improvements average 7.30%). These affect testability and consistency, but neither reduces the claimed result to its inputs by construction. I therefore find no significant circularity; the score reflects only minor, non-load-bearing self-citation overlap.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The central claim rests on a large number of unstated hyperparameters (restart probability, sample counts, combination weights, learning rates) and on qualitative assumptions about the quality of mined rules and the meaningfulness of treating all graph edges as user-item interactions. The rule-mining module is not specified, so the 'rules' that drive the random walks are the least auditable part of the pipeline.

free parameters (6)

restart probability alpha = not reported
Controls the length of forward and backward random walks in Equations 2 and 3; the value is not given in Section 4.2.
sampled node-attribute pairs n_r = not reported
Determines the size of the sampled pair sets S_f and S_b in Section 4.2; no value is specified.
embedding dimension k = 64
Set to 64 in Section 5.1, with k/2 allocated to forward and backward embeddings; impacts representation capacity.
learning rate = searched over 0.01, 0.001, 0.0001
Section 5.1 reports a search but not the selected values for each dataset.
L2 regularization coefficient lambda = searched in [10^-5, 10^-2]
Section 5.1 reports a search range but not the selected values.
combination weights alpha and beta in Eq. 5 = not reported
The unified transformation in Equation 5 uses these weights to blend forward and backward embeddings; their values or tuning procedure are not described.

assumptions (5)

standard math Random walks with restart produce valid pointwise mutual information estimates for node-attribute affinities.
Equations 2 and 3 define shifted PMI scores from RWR frequencies, following the argument in PANE [13,14].
domain assumption Mined association rules from the property graph are reliable and transferable to user-item preference prediction.
The motivation in Section 1 assumes rules like V2.MAJOR -> V1.GENRE capture genuine preference dependencies; no validation of rule quality is provided.
domain assumption All edges in the experimental datasets can be treated as user-item interactions.
Section 5.1 lists Facebook, BlogCatalog, Flickr, Citeseer without explaining how user-item interactions are constructed from each graph, yet the problem formulation (Section 3.2) requires such interactions.
ad hoc to paper The rule jump semantics are well-defined and beneficial.
Section 4.2 introduces LHS propagation and rule jumps only informally; no formal definition or analysis is given.
domain assumption LightGCN's architecture remains effective when its input embeddings are replaced by the combined rule-based embeddings.
Section 4.3 and 5.1 adopt LightGCN [19] as the recommendation backbone without modification or justification beyond the original paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of RAE: A Rule-Driven Approach for Attribute Embedding in Property Graph Recommendation." pith.science (2026). https://pith.science/paper/JLXYE4PK

@misc{pith2026250608314,
  author       = {Pith},
  title        = {Pith review of: RAE: A Rule-Driven Approach for Attribute Embedding in Property Graph Recommendation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JLXYE4PK}},
  note         = {Machine review of arXiv:2506.08314}
}

read the original abstract

Recommendation systems are crucial in modern applications to enhance the user experience and drive business conversion rates through personalization. However, insufficient utilization of attribute information within the property graph remains a significant challenge. Most existing graph convolutional network (GCN) models do not consider attribute information, and those that do often employ a simplified triple format <users, items, attributes>, which fails to fully exploit the rich semantic structures of property graphs necessary for effective recommendations. To overcome these limitations, we introduce Rule-Driven Approach for Attribute Embedding (RAE), a novel methodology that enhances recommendation performance by effectively mining and utilizing semantic rules from property graphs. RAE applies a rule-mining process to extract meaningful rules that guide random walks in generating enriched attribute embeddings. These enriched embeddings are subsequently integrated into GCNs, surpassing conventional triple-based embedding techniques. We evaluate RAE on real-world datasets (e.g., Blogcatalog and Flickr) and demonstrate that RAE achieves an average improvement of 10.6% in both Recall@20 and NDCG@20 compared to state-of-the-art baselines, indicating superior relevance coverage and ranking rationality in top-20 recommendations. Additionally, RAE exhibits enhanced robustness against data sparsity and the attribute missingness problem. Our novel approach underscores the significant performance gains achieved in recommendation systems by fully leveraging attribute information within property graphs, enhancing both effectiveness and reliability.

Figures

Figures reproduced from arXiv: 2506.08314 by the authors.

**Figure 1.** Illustration of motivation of the proposed RAE 1 Introduction Recommendation systems play an essential role in modern online applications, enhancing user experience by providing personalized content and services. These systems are crucial in e-commerce, social media, and entertainment platforms, helping users discover relevant items and mitigate information overload. Traditional recommendation algorithms primarily … view at source ↗

**Figure 2.** A Graph Pattern and Attribute-Extended Graph G ′ Rules in Graph: A rule in a graph φ is defined as: Q[¯x](X → Y ), where Q[¯x] denotes a graph pattern, and X and Y represent conjunctions of literals within Q[¯x]. We refer to Q[¯x] as the pattern and X → Y as the dependency of φ. For instance, consider a rule ( [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

**Figure 3.** (a)-(b) Perfomance comparison of the sparsity distribution of the user on different datasets. (c)-(d) Impact of attributes missing with attribute removed randomly in ratio of 0, 0.2, 0.4, 0.8. k is the ratio of removed attributes in each dataset. 5.3 Effects of Data Sparsity Attribute information plays a crucial role in addressing data sparsity issues in recommendation systems. To examine the effectiveness of RAE u… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

37 extracted references · 19 canonical work pages

[12]

Fan, W., Liu, M., Liu, S., Tian, C.: Capturing More Associations by Ref- erencing External Graphs. Proc. VLDB Endow. 17(6), 1173–1186 (2024). https://doi.org/10.14778/3648160.3648162

arXiv 2024
[1]

In: Proceedings of the European Conference on Machine Learning and Prin- ciples and Practice of Knowledge Discovery in Databases (ECML/PKDD), pp

Berlingerio, M., Bonchi, F., Bringmann, B., Gionis, A.: Mining Graph Evolution Rules. In: Proceedings of the European Conference on Machine Learning and Prin- ciples and Practice of Knowledge Discovery in Databases (ECML/PKDD), pp. 115–130. (2009)

work page 2009
[2]

1191–1196

Scharwächter, E., Müller, E., Donges, J.F., Hassani, M., Seidl, T.: Detecting Change Processes in Dynamic Networks by Frequent Graph Evolution Rule Min- ing.In:ProceedingsoftheIEEEInternationalConferenceonDataMining(ICDM), pp. 1191–1196. (2016)

work page 2016
[3]

In: Proceedings of the ACM International Conference on Infor- mation and Knowledge Management (CIKM), pp

Leung, C.W., Lim, E.-P., Lo, D., Weng, J.: Mining interesting link formation rules in social networks. In: Proceedings of the ACM International Conference on Infor- mation and Knowledge Management (CIKM), pp. 209–218. (2010)

work page 2010
[4]

In: Proceedings of the Interna- tional Conference on Learning Representations (ICLR)

Qu, M., Chen, J., Xhonneux, L.-P.A.C., Bengio, Y., Tang, J.: RNNLogic: Learning Logic Rules for Reasoning on Knowledge Graphs. In: Proceedings of the Interna- tional Conference on Learning Representations (ICLR). (2021)

work page 2021
[5]

In: Pro- ceedings of the Information Technologies – Applications and Theory (ITAT), pp

Vaculík, K.: A Versatile Algorithm for Predictive Graph Rule Mining. In: Pro- ceedings of the Information Technologies – Applications and Theory (ITAT), pp. 51–58. (2015)

work page 2015
[6]

In: Proceed- ings of the IEEE International Conference on Data Mining (ICDM), pp

Yan, X., Han, J.: gSpan: Graph-Based Substructure Pattern Mining. In: Proceed- ings of the IEEE International Conference on Data Mining (ICDM), pp. 721–724. (2002)

work page 2002
[7]

Fan, W., Fu, W., Jin, R., Liu, M., Lu, P., Tian, C.: Making It Tractable to Catch Duplicates and Conflicts in Graphs. Proc. ACM Manag. Data 1(1), 86:1–86:28 (2023)

work page 2023

Show all 37 references

[8]

Fan, W., Fu, W., Jin, R., Lu, P., Tian, C.: Discovering Association Rules from Big Graphs. Proc. VLDB Endow.15(7), 1479–1492 (2022)

2022
[9]

ACM Trans

Fan, W., Hu, C., Liu, X., Lu, P.: Discovering Graph Functional Dependencies. ACM Trans. Database Syst.45(3), 15:1–15:42 (2020)

2020
[10]

Kwashie, S., Liu, J., Li, J., Liu, L., Stumptner, M., Yang, L.: Certus: An Effective Entity Resolution Approach with Graph Differential Dependencies (GDDs). Proc. VLDB Endow. 12(6), 653–666 (2019)

2019
[11]

In: Zhang, F., Wang, H., Barhamgi, M., Chen, L., Zhou, R

Zhou, G., Kwashie, S., Zhang, Y., Bewong, M., Nofong, V.M., Hu, J., Cheng, D., He, K., Liu, S., Feng, Z.: FastAGEDs: Fast Approximate Graph Entity Dependency Discovery. In: Zhang, F., Wang, H., Barhamgi, M., Chen, L., Zhou, R. (eds.) Web Information Systems Engineering – WISE ...

2023 doi
[13]

Proceedings of the VLDB Endowment (PVLDB), 14(1), 37–49 (2020)

Yang, R., Shi, J., Xiao, X., Yang, Y., Liu, J., Bhowmick, S.S.: Scaling Attributed Network Embedding to Massive Graphs. Proceedings of the VLDB Endowment (PVLDB), 14(1), 37–49 (2020). https://doi.org/10.14778/3421424.3421430

2020
[14]

The VLDB Journal,32(6), 1237–1262 (2023)

Yang, R., Shi, J., Xiao, X., Yang, Y., Bhowmick, S.S., Liu, J.: PANE: Scalable and Effective Attributed Network Embedding. The VLDB Journal,32(6), 1237–1262 (2023). Springer, Berlin, Heidelberg. https://doi.org/10.1007/s00778-023-00790-4

2023 doi
[15]

Journal of Chemometrics23, 393–405 (2009)

Comon, P., Luciani, X., De Almeida, A.L.F.: Tensor Decompositions, Alternat- ing Least Squares and other Tales. Journal of Chemometrics23, 393–405 (2009). https://doi.org/10.1002/cem.1236

2009 doi
[16]

IEEE Trans

Wang, J., Qu, X., Bai, J., Li, Z., Zhang, J., Gao, J.: Sages: Scalable attributed graph embedding with sampling for unsupervised learning. IEEE Trans. Knowl. Data Eng., 1–1 (2022). https://doi.org/10.1109/TKDE.2022.3148272

2022
[17]

In: Proceedings of the 28th International Joint Conference on Artificial Intelligence (IJCAI’19), pp

Yang, H., Pan, S., Chen, L., Zhou, C., Zhang, P.: Low-bit quantization for at- tributed network representation learning. In: Proceedings of the 28th International Joint Conference on Artificial Intelligence (IJCAI’19), pp. 4047–4053. AAAI Press (2019)

2019
[18]

In: IEEE International Conference on Data Mining (ICDM), pp

Yang, H., Pan, S., Zhang, P., Chen, L., Lian, D., Zhang, C.: Binarized attributed network embedding. In: IEEE International Conference on Data Mining (ICDM), pp. 1476–1481. (2018). https://doi.org/10.1109/ICDM.2018.8626170

2018
[19]

In: Proceed- ings of the 43rd International ACM SIGIR Conference on Research and Develop- ment in Information Retrieval (SIGIR ’20), pp

He, X., Deng, K., Wang, X., Li, Y., Zhang, Y., Wang, M.: LightGCN: Simplify- ing and Powering Graph Convolution Network for Recommendation. In: Proceed- ings of the 43rd International ACM SIGIR Conference on Research and Develop- ment in Information Retrieval (SIGIR ’20), pp. ...

2020
[20]

In: The Web Conference 2021 (WWW), pp

Liu, F., Cheng, Z., Zhu, L., Gao, Z., Nie, L.: Interest-aware Message-Passing GCN for Recommendation. In: The Web Conference 2021 (WWW), pp. 1296–1305. ACM, New York (2021). https://doi.org/10.1145/3442381.3449986

2021
[21]

In: Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR), pp

Wang, X., He, X., Wang, M., Feng, F., Chua, T.-S.: Neural Graph Collaborative Filtering. In: Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR), pp. 165–174. ACM, New York (2019). https://doi.org/10.1145/3331...

2019
[22]

In: KDD, London, UK, pp

Ying, R., He, R., Chen, K., Eksombatchai, P., Hamilton, W.L., Leskovec, J.: Graph Convolutional Neural Networks for Web-Scale Recommender Systems. In: KDD, London, UK, pp. 974–983. (2018). https://doi.org/10.1145/3219819.3219890

2018
[23]

IEEE Trans

Liu, F., Cheng, Z., Zhu, L., Liu, C., Nie, L.: An Attribute-Aware Attentive GCN Model for Attribute Missing in Recommendation. IEEE Trans. Knowl. Data Eng. 34(9), 4077–4088 (2022). https://doi.org/10.1109/TKDE.2020.3040772

2022
[24]

IEEE Trans

Yue, G., Xiao, R., Zhao, Z., Li, C.: AF-GCN: Attribute-Fusing Graph Convolu- tion Network for Recommendation. IEEE Trans. Big Data9(2), 597–607 (2023). https://doi.org/10.1109/TBDATA.2022.3192598

2023
[25]

In: UAI, pp

Rendle, S., Freudenthaler, C., Gantner, Z., Schmidt-Thieme, L.: BPR: Bayesian personalized ranking from implicit feedback. In: UAI, pp. 452–461. (2009)

2009
[26]

In: Proc

Jeh, G., Widom, J.: Scaling personalized web search. In: Proc. 12th International Conference on World Wide Web (WWW), pp. 271–279. ACM, New York (2003). https://doi.org/10.1145/775152.775191

2003
[27]

In: 6th International Conference on Data Mining (ICDM), pp

Tong, H., Faloutsos, C., Pan, J.-Y.: Fast Random Walk with Restart and Its Ap- plications. In: 6th International Conference on Data Mining (ICDM), pp. 613–622. IEEE Computer Society, USA (2006). https://doi.org/10.1109/ICDM.2006.70 18 S. Zhao et al

2006 doi
[28]

Liu, D., Kwashie, S., Zhang, Y., Zhou, G., Bewong, M., Wu, X., Guo, X., He, K., Feng, Z.: An efficient approach for discovering Graph Entity Dependencies (GEDs). Inf. Syst. 125, 102421 (2024). https://doi.org/10.1016/j.is.2024.102421

2024
[29]

In: Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR ’17), pp

He,X.,Chua,T.-S.:NeuralFactorizationMachinesforSparsePredictiveAnalytics. In: Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR ’17), pp. 355–364. ACM, New York, NY, USA (2017). https://doi.org/10.1145/307713...

2017
[30]

ACM Trans

Deshpande, M., Karypis, G.: Item-based top-N recommendation algorithms. ACM Trans. Inf. Syst.22(1), 143–177 (2004). https://doi.org/10.1145/963770.963776

2004
[31]

Hu, J., Bewong, M., Kwashie, S., Zhang, Y., Nofong, V., Wondoh, J., Feng, Z.: When GDD meets GNN: A knowledge-driven neural connection for ef- fective entity resolution in property graphs. Inf. Syst. 132, 102551 (2025). https://doi.org/10.1016/j.is.2025.102551

2025
[32]

In: The Eleventh International Conference on Learning Representations (ICLR)

Cai, X., Huang, C., Xia, L., Ren, X.: LightGCL: Simple Yet Effective Graph Con- trastive Learning for Recommendation. In: The Eleventh International Conference on Learning Representations (ICLR). (2023)

2023
[33]

In: Proceedings of the 17th ACM International Conference on Web Search and Data Mining (WSDM), pp

Ren, X., Xia, L., Yang, Y., Wei, W., Wang, T., Cai, X., Huang, C.: SSLRec: A Self-Supervised Learning Framework for Recommendation. In: Proceedings of the 17th ACM International Conference on Web Search and Data Mining (WSDM), pp. 567–575. Mérida, Mexico (2024). https://doi.or...

2024
[34]

In: Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR), pp

He, W., Sun, G., Lu, J., Fang, X.S.: Candidate-aware Graph Contrastive Learn- ing for Recommendation. In: Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR), pp. 1670–1679. Taipei, Taiwan (2023). https://doi....

2023
[35]

In: Proceedings of the 27th ACM International Conference on Informa- tion and Knowledge Management (CIKM), pp

Wang, H., Zhang, F., Wang, J., Zhao, M., Li, W., Xie, X., Guo, M.: RippleNet: Propagating User Preferences on the Knowledge Graph for Recommender Sys- tems. In: Proceedings of the 27th ACM International Conference on Informa- tion and Knowledge Management (CIKM), pp. 417–426. ...

2018
[36]

In: The World Wide Web Conference (WWW), pp

Wang, H., Zhao, M., Xie, X., Li, W., Guo, M.: Knowledge Graph Con- volutional Networks for Recommender Systems. In: The World Wide Web Conference (WWW), pp. 3307–3313. San Francisco, CA, USA (2019). https://doi.org/10.1145/3308558.3313417

2019
[37]

In: The World Wide Web Conference (WWW), pp

Ma, W., Zhang, M., Cao, Y., Jin, W., Wang, C., Liu, Y., Ma, S., Ren, X.: Jointly Learning Explainable Rules for Recommendation with Knowledge Graph. In: The World Wide Web Conference (WWW), pp. 1210–1221. San Francisco, CA, USA (2019). https://doi.org/10.1145/3308558.3313607

2019

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [12]

Fan, W., Liu, M., Liu, S., Tian, C.: Capturing More Associations by Ref- erencing External Graphs. Proc. VLDB Endow. 17(6), 1173–1186 (2024). https://doi.org/10.14778/3648160.3648162

arXiv 2024

[2] [1]

In: Proceedings of the European Conference on Machine Learning and Prin- ciples and Practice of Knowledge Discovery in Databases (ECML/PKDD), pp

Berlingerio, M., Bonchi, F., Bringmann, B., Gionis, A.: Mining Graph Evolution Rules. In: Proceedings of the European Conference on Machine Learning and Prin- ciples and Practice of Knowledge Discovery in Databases (ECML/PKDD), pp. 115–130. (2009)

work page 2009

[3] [2]

1191–1196

Scharwächter, E., Müller, E., Donges, J.F., Hassani, M., Seidl, T.: Detecting Change Processes in Dynamic Networks by Frequent Graph Evolution Rule Min- ing.In:ProceedingsoftheIEEEInternationalConferenceonDataMining(ICDM), pp. 1191–1196. (2016)

work page 2016

[4] [3]

In: Proceedings of the ACM International Conference on Infor- mation and Knowledge Management (CIKM), pp

Leung, C.W., Lim, E.-P., Lo, D., Weng, J.: Mining interesting link formation rules in social networks. In: Proceedings of the ACM International Conference on Infor- mation and Knowledge Management (CIKM), pp. 209–218. (2010)

work page 2010

[5] [4]

In: Proceedings of the Interna- tional Conference on Learning Representations (ICLR)

Qu, M., Chen, J., Xhonneux, L.-P.A.C., Bengio, Y., Tang, J.: RNNLogic: Learning Logic Rules for Reasoning on Knowledge Graphs. In: Proceedings of the Interna- tional Conference on Learning Representations (ICLR). (2021)

work page 2021

[6] [5]

In: Pro- ceedings of the Information Technologies – Applications and Theory (ITAT), pp

Vaculík, K.: A Versatile Algorithm for Predictive Graph Rule Mining. In: Pro- ceedings of the Information Technologies – Applications and Theory (ITAT), pp. 51–58. (2015)

work page 2015

[7] [6]

In: Proceed- ings of the IEEE International Conference on Data Mining (ICDM), pp

Yan, X., Han, J.: gSpan: Graph-Based Substructure Pattern Mining. In: Proceed- ings of the IEEE International Conference on Data Mining (ICDM), pp. 721–724. (2002)

work page 2002

[8] [7]

Fan, W., Fu, W., Jin, R., Liu, M., Lu, P., Tian, C.: Making It Tractable to Catch Duplicates and Conflicts in Graphs. Proc. ACM Manag. Data 1(1), 86:1–86:28 (2023)

work page 2023