Pith. sign in

REVIEW 3 major objections 4 minor 34 references

Context Pooling: Query-specific Graph Pooling for Generic Inductive Link Prediction in Knowledge Graphs

T0 review · 3 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Context Pooling claims that query-specific graph pooling based on neighborhood precision and recall improves inductive link prediction in knowledge graphs, achieving state-of-the-art results in 42 of 48 settings.

desk verdict A genuinely new pooling layer for KG link prediction with clean integration, but the SOTA claim is shaky because the baseline protocol isn't confirmed and the win count doesn't add up. read the letter →

arxiv 2507.07595 v1 pith:DSDOTRND submitted 2025-07-10 cs.AI cs.LG

Zhixiang Su , Di Wang , Chunyan Miao This is my paper
classification cs.AIcs.LG
keywords KnowledgeGraphsLinkPredictionGraphNeuralNetworksPoolingInductiveQuery-specificNeighborhoodPrecisionContextNeighborFamily
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

Context Pooling attempts to establish that GNN-based link prediction in knowledge graphs is better when the model does not aggregate all neighbors, but only neighbors judged logically relevant to the query relation. The paper defines two statistics, neighborhood precision and neighborhood recall, computed from relation co-occurrence in the training graph, and uses them to build a query-specific context graph at every aggregation hop. The method is generic: wrapping RED-GNN and NBFNet with Context Pooling produces state-of-the-art results in 42 of 48 transductive and inductive settings across WN18RR, FB15k-237, and NELL-995. If the claim holds, graph pooling becomes a cheap, training-free plug-in that directly targets the recent observation that vanilla neighbor aggregation contributes little to KG link prediction.

What carries the argument

The central object is the Context Neighbor Family (CNF) together with the two scores that define it. For a query relation $r$ and a candidate set of neighboring relations $\mathrm{NR}'$, neighborhood precision is $\mathrm{Rel}_{\mathrm{pre}}(\mathrm{NR}', r) = |\{e : \mathrm{NR}' \cup \{r\} \subseteq \mathrm{NR}_e\}| \,/\, |\{e : \mathrm{NR}' \subseteq \mathrm{NR}_e\}|$, and neighborhood recall is $\mathrm{Rel}_{\mathrm{rec}}(\mathrm{NR}', r) = |\{e : \mathrm{NR}' \cup \{r\} \subseteq \mathrm{NR}_e\}| \,/\, |\{e : r \in \mathrm{NR}_e\}|$. CNF collects all relation sets whose combined relevance score clears a threshold. Because the scores depend only on relations, they transfer to unseen entities. The optimized Algorithm 3 invokes Assumption 1, pairwise and conditional independence given the query relation, so that combined scores factor into products of per-relation scores; the chosen neighbors at each hop are the intersection of the current entity's neighboring relations with the stored high-scoring set. This per-hop selection acts as the score generator, node selector, and graph coarsening of a node-drop pooling operation specialized to heterogeneous knowledge graphs.

What would settle it

Construct or select a KG where the co-occurrence of neighboring relations is strongly correlated with the query relation, for instance a domain where the relation Profession usually co-occurs with Award only when Company is also present, and run RED-GNN+CP with Algorithm 3 versus plain RED-GNN. If the pooled model's MRR drops below the base model while the unoptimized Algorithm 1 does not drop, the independence assumption is the load-bearing approximation.

Watch

Extended reading notes

Core claim

The paper's central discovery is that the logically relevant neighbors of a query can be identified without entity embeddings or a schema, using only counts of relation co-occurrences. For a query relation $r$, the context neighbor family $\mathrm{CNF}(r)$ is the collection of relation sets $\mathrm{NR}'$ whose neighborhood precision and neighborhood recall exceed a threshold. Neighborhood precision is the fraction of entities whose neighborhood contains $\mathrm{NR}'$ that also contain $r$; neighborhood recall is the fraction of entities containing $r$ that also contain $\mathrm{NR}'$. Because these scores are computed from the training graph and depend only on relations, they transfer to unseen entities. The optimized version of the method (Algorithm 3) assumes that any two non-redundant neighboring relations are independent given the query relation; then combined scores factorize, and the best neighbor set for a query is simply the intersection of the current entity's neighboring relations with the stored high-scoring relations. The paper reports that this query-specific pooling lifts RED-GNN and NBFNet to state-of-the-art MRR and Hit@1 in 42 of 48 settings, with paired t-test $p$-values below $10^{-3}$ for transductive MRR and Hit@1 and for inductive MRR, and $2 \times 10^{-3}$ for inductive Hit@1.

Load-bearing premise

All reported results use Algorithm 3, whose correctness relies on Assumption 1: that any two non-redundant neighboring relations occur independently of each other, and independently given the query relation; the paper's own ablation shows this assumption is approximate, since exact search reaches higher MRR on FB15k-237-V1.

Editorial extensions

If this is right

  • RED-GNN+CP and NBFNet+CP can be dropped into existing GNN pipelines with only two thresholds to set; on 88 of 96 transductive and inductive settings the pooled version beats the base model.
  • Pooling cost is $O(|R_G|^2)$ for training and $O(|\mathrm{CNF}'|)$ per query, so the scalability bottleneck remains the GNN itself.
  • Because the selection is relation-based, the same context neighbor family learned on the training graph applies to test entities that were never seen at training time.
  • On WN18RR-V2 in the transductive setting the gain over NBFNet is 11.7% in MRR and 19.4% in Hit@1; on NELL-995-V4 in the inductive setting the gain over RED-GNN is 11.7% in MRR and 16.8% in Hit@1.

Reading between the lines

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

  • Inference: the same relation-co-occurrence statistics could be used to pre-compute a per-relation support graph before training, shrinking the effective graph for memory-bound GNNs; the paper does not test this.
  • Inference: because the pooling layer is unsupervised and entity-free, it should transfer to other relation-heavy graphs beyond knowledge graphs, such as heterogeneous social or biological networks with typed edges.
  • Inference: the ablation suggests a tunable trade-off — on small relation sets the exact power-set search is preferable (FB15k-237-V1 MRR 0.396 versus 0.383 for the optimized version), while on large relation sets only the factorized version is feasible; a hybrid that uses exact search on small relation neighborhoods could recover the lost accuracy.
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

3 major / 4 minor

Summary. The paper introduces Context Pooling (CP), a training-free, query-specific graph pooling method for inductive knowledge-graph link prediction. CP scores neighboring relations by empirical neighborhood precision and recall computed from the training graph, prunes neighbors that fall below thresholds, and feeds the resulting query-specific subgraph alongside the original graph into a GNN such as NBFNet or RED-GNN. The authors provide an exact power-set formulation (Algorithm 1), an optimized formulation (Algorithm 3) under a conditional-independence assumption, a proof that the optimized relevance scores are proportional to products of singleton scores, and experiments on three datasets with four versions each in transductive and inductive settings. They report that CP improves the two base models and achieves SOTA performance in 42 of 48 settings.

Significance. If the SOTA claim can be supported under a uniform evaluation protocol, CP is a simple and generic plug-in component with inductive capability, and the paper would make a useful contribution to KG link prediction. Strengths include the clear formalization of Definitions 5 and 6, the explicit independence assumption rather than a hidden one, the released code, and the paired t-test evidence that CP improves NBFNet and RED-GNN. The co-occurrence statistics are computed from the training graph rather than from test labels, so the method is not circular. The main reservations are the fairness of the baseline comparison and the limited validation of the independence assumption used by the optimized algorithm.

major comments (3)
  1. [Section 5, Tables 1-2] The SOTA comparison may not be apples-to-apples. Section 5 states that the authors use the all-entities-as-negatives protocol from RED-GNN [31], unlike GRAIL's 49-negative protocol, but the manuscript never states that the rows for Neural LP, DRUM, CompGCN, AnyBURL, and GRAIL were recomputed under this protocol. If those rows are taken from prior papers that used an easier negative-sampling setup, their MRR and Hits@1 values are inflated relative to the CP models, and the headline "42 out of 48 settings" is not established. The paired t-test in Section 5.1.3 only compares CP against NBFNet and RED-GNN, so it does not validate superiority over the other baselines. Please recompute all baselines under the same protocol or explicitly restrict the SOTA claim to a fair comparison.
  2. [Section 4.3, Table 4] Assumption 1 is load-bearing and only weakly validated. All main results use Algorithm 3, whose correctness rests on the conditional-independence assumption. The ablation in Section 5.3 (Table 4) is the only evidence for this assumption, and it is mixed: on FB15k-237-V1 the unoptimized Algorithm 1 gives MRR 0.396 vs 0.383 for the optimized algorithm (+3.4%) and Hit@1 0.332 vs 0.316 (+5.1%), while on WN18RR-V1 the optimized version is slightly better. The paper attributes the WN18RR-V1 result to noise, but the FB15k-237-V1 gap shows that the independence assumption can change the selected neighbors and measurably reduce accuracy. Moreover, the unoptimized algorithm in the ablation is not exactly Algorithm 1, because Section 5.3 restricts the relation count to [4,6). Please validate the assumption on more splits/datasets, report results for Algorithm 1 where feasible, or provide a theoretical bound on the error introduced by the approximation.
  3. [Section 4.3, Algorithm 3, Section 5] The scoring function and thresholds are under-specified. Definition 4 defines CNF via a generic Rel(), and Definitions 5-6 define two separate metrics, neighborhood precision and neighborhood recall. Algorithm 3 line 12 tests Rel(r', r_i) > R0' but the paper never states whether the threshold is applied to precision, to recall, or to a combination (e.g., product or minimum), nor how the two thresholds are chosen. Section 5 only says the thresholds "are set within the range [10^-5, 10^-1]", which spans four orders of magnitude and is not a reproducible specification. Please state the exact scoring rule and per-dataset threshold values, or the criterion used to select them.
minor comments (4)
  1. [Sections 5.1.1 and 5.1.2] The counts in the text do not match the tables: after stating RED-GNN+CP is best in 10/12 transductive settings, the text says "In the remaining 3 of 4 settings" instead of 2, and after 8/12 inductive settings it says "remaining 5 of 8 settings" instead of 4. Please correct these counts.
  2. [Section 2.2 / Section 1] The novelty claim that Context Pooling is the first graph pooling method for KGs would be stronger with a brief comparison to prior work on selectively aggregating neighbors in knowledge graphs; currently the related work only covers homogeneous-graph pooling.
  3. [Table 3] In Table 3, the '-1' suffix used for inverse relations is not explained in the caption; please clarify the notation and specify whether the precision and recall columns are separate neighbor sets or a combined set.
  4. [Section 4.3] The phrase "Markov chain-inspired approach" is misleading because Theorem 1 follows directly from the independence assumption and does not require Markov chain machinery; consider removing the reference or rewording.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Context Pooling's relevance statistics are fitted on training relational co-occurrence, while link-prediction accuracy is measured on held-out queries by external GNN scorers.

full rationale

The derivation chain is self-contained. Neighborhood precision and recall (Definitions 5 and 6) are computed from co-occurrence statistics over the training graph and are used only to select query-specific neighbor relations (Algorithms 1-3); the link-prediction score is produced by the NBFNet/RED-GNN scoring function on the pooled graph (Equations 7-11), so the reported MRR and Hits@1 are not the same quantities being fitted. Theorem 1 is derived explicitly from Assumption 1, which is stated as an assumption and not as the target result; the ablation in Table 4 even shows that the optimized approximation can lose accuracy on FB15k-237-V1, so the theorem is not vacuous and the pooling choice is not forced by construction. The only self-citation is reference [20], used to define inductive link prediction as background notation and not as evidence for the central pooling claim. The skeptical concern about baseline negative-sampling protocols is an experimental-fairness issue, not a circularity of the derivation chain. No step reduces a predicted quantity to its own input by construction.

Assumptions & free parameters 1 free parameters · 2 assumptions · 0 invented entities

The method introduces one hyperparameter (threshold) and two statistical assumptions (independence, transfer of relation statistics). No invented entities or forces.

free parameters (1)
  • Threshold R0' for CNF' (neighborhood precision/recall) = not reported; grid range [1e-5, 1e-1]
    The paper states thresholds are set within [1e-5, 1e-1] but does not report per-dataset values or validation split, so the selection procedure is unclear.
assumptions (2)
  • ad hoc to paper Conditional independence of neighboring relations (Assumption 1): for any query relation r, occurrences of any two non-redundant neighbor relations are independent and conditionally independent given r.
    Introduced to reduce the exponential power-set search in Algorithm 1 to per-relation scoring in Algorithm 3. The ablation on FB15k-237-V1 shows the full method gives MRR 0.396 vs 0.383 for the optimized one, so the assumption is materially approximate.
  • domain assumption Relation co-occurrence statistics measured on the training graph transfer to test entities in inductive settings.
    In the inductive setting (Definition 1) no entity overlaps, so the CNF built on G_train must generalize to unseen entities; the paper only validates this empirically on three datasets.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Context Pooling: Query-specific Graph Pooling for Generic Inductive Link Prediction in Knowledge Graphs." pith.science (2026). https://pith.science/paper/DSDOTRND

@misc{pith2026250707595,
  author       = {Pith},
  title        = {Pith review of: Context Pooling: Query-specific Graph Pooling for Generic Inductive Link Prediction in Knowledge Graphs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DSDOTRND}},
  note         = {Machine review of arXiv:2507.07595}
}
read the original abstract

Recent investigations on the effectiveness of Graph Neural Network (GNN)-based models for link prediction in Knowledge Graphs (KGs) show that vanilla aggregation does not significantly impact the model performance. In this paper, we introduce a novel method, named Context Pooling, to enhance GNN-based models' efficacy for link predictions in KGs. To our best of knowledge, Context Pooling is the first methodology that applies graph pooling in KGs. Additionally, Context Pooling is first-of-its-kind to enable the generation of query-specific graphs for inductive settings, where testing entities are unseen during training. Specifically, we devise two metrics, namely neighborhood precision and neighborhood recall, to assess the neighbors' logical relevance regarding the given queries, thereby enabling the subsequent comprehensive identification of only the logically relevant neighbors for link prediction. Our method is generic and assessed by being applied to two state-of-the-art (SOTA) models on three public transductive and inductive datasets, achieving SOTA performance in 42 out of 48 settings.

Figures

Figures reproduced from arXiv: 2507.07595 by the authors.

Figure 1
Figure 1. A KG comprising 2-hop neighbors of MorganFreeman. A KG, represented as 𝐺(𝐸𝐺, 𝑅𝐺 ) = {(ℎ𝑖 , 𝑟𝑖 , 𝑡𝑖)|𝑖 = 1, 2, 3, ...,𝑚}, is a heterogeneous graph of interconnected entities (heads and tails, ℎ𝑖 , 𝑡𝑖 ∈ 𝐸𝐺 ) and relations (𝑟𝑖 ∈ 𝑅𝐺 ). In this context, link prediction is a critical process making inference of the missing entity in a given query triple. This involves the prediction of either the tail entity 𝑡 given (ℎ, 𝑟… view at source ↗
Figure 2
Figure 2. An illustration of Context Pooling GNN architecture. [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

34 extracted references · 24 canonical work pages

  1. [31]

    Yongqi Zhang and Quanming Yao. 2022. Knowledge graph reasoning with relational digraph. InProceedings of the ACM web conference. 912–924

    work page 2022

  2. [1]

    Jinheon Baek, Minki Kang, and Sung Ju Hwang. 2021. Accurate learning of graph representations with graph multiset pooling.arXiv preprint arXiv:2102.11533 (2021)

    arXiv 2021

  3. [2]

    Chen Cai, Dingkang Wang, and Yusu Wang. 2021. Graph coarsening with neural networks.arXiv preprint arXiv:2102.01350(2021)

    arXiv 2021

  4. [3]

    Tim Dettmers, Pasquale Minervini, Pontus Stenetorp, and Sebastian Riedel. 2018. Convolutional 2d knowledge graph embeddings. InProceedings of the AAAI conference on artificial intelligence, Vol. 32

    work page 2018

  5. [4]

    David K Duvenaud, Dougal Maclaurin, Jorge Iparraguirre, Rafael Bombarell, Timothy Hirzel, Alán Aspuru-Guzik, and Ryan P Adams. 2015. Convolutional networks on graphs for learning molecular fingerprints.Advances in neural information processing systems28 (2015)

    work page 2015

  6. [5]

    Xing Gao, Wenrui Dai, Chenglin Li, Hongkai Xiong, and Pascal Frossard. 2021. ipool—information-based pooling in hierarchical graph neural networks.IEEE Transactions on Neural Networks and Learning Systems33, 9 (2021), 5032–5044

    work page 2021

  7. [6]

    Xiao Huang, Jingyuan Zhang, Dingcheng Li, and Ping Li. 2019. Knowledge graph embedding based question answering. InProceedings of the twelfth ACM international conference on web search and data mining. 105–113

    work page 2019

  8. [7]

    Markus Krötzsch, Maximilian Marx, Ana Ozaki, and Veronika Thost. 2018. At- tributed description logics: Reasoning on knowledge graphs.. InInternational Joint Conferences on Artificial Intelligence. 5309–5313

    work page 2018

Show all 34 references
  1. [8]

    Juanhui Li, Harry Shomer, Jiayuan Ding, Yiqi Wang, Yao Ma, Neil Shah, Jil- iang Tang, and Dawei Yin. 2022. Are graph neural networks really helpful for knowledge graph completion?arXiv preprint arXiv:2205.10652(2022)

    2022 arXiv

  2. [9]

    Linfeng Li, Peng Wang, Jun Yan, Yao Wang, Simin Li, Jinpeng Jiang, Zhe Sun, Buzhou Tang, Tsung-Hui Chang, Shenghui Wang, et al. 2020. Real-world data medical knowledge graph: construction and applications.Artificial intelligence in medicine103 (2020), 101817

    2020

  3. [10]

    Maosen Li, Siheng Chen, Ya Zhang, and Ivor Tsang. 2020. Graph cross networks with vertex infomax pooling.Advances in Neural Information Processing Systems 33 (2020), 14093–14105

    2020

  4. [11]

    Xuan Lin, Zhe Quan, Zhi-Jie Wang, Tengfei Ma, and Xiangxiang Zeng. 2020. KGNN: Knowledge Graph Neural Network for Drug-Drug Interaction Prediction. International Joint Conferences on Artificial Intelligence(2020), 2739–2745

    2020

  5. [12]

    Chuang Liu, Yibing Zhan, Jia Wu, Chang Li, Bo Du, Wenbin Hu, Tongliang Liu, and Dacheng Tao. 2022. Graph pooling for graph neural networks: Progress, challenges, and opportunities.arXiv preprint arXiv:2204.07321(2022)

    2022 arXiv

  6. [13]

    Farzaneh Mahdisoltani, Joanna Biega, and Fabian M Suchanek. 2013. Yago3: A knowledge base from multilingual wikipedias. InCIDR

    2013

  7. [14]

    Christian Meilicke, Melisachew Wudage Chekol, Manuel Fink, and Heiner Stuck- enschmidt. 2020. Reinforced anytime bottom up rule learning for knowledge graph completion.arXiv preprint arXiv:2004.04412(2020)

    2020 arXiv

  8. [15]

    Christian Meilicke, Manuel Fink, Yanjie Wang, Daniel Ruffinelli, Rainer Gemulla, and Heiner Stuckenschmidt. 2018. Fine-grained evaluation of rule-and embedding-based systems for knowledge graph completion. InProceedings of International Semantic Web Conference. Springer, 3–20

    2018

  9. [16]

    1998.Markov chains

    James R Norris. 1998.Markov chains. Number 2. Cambridge university press

    1998

  10. [17]

    Meng Qu and Jian Tang. 2019. Probabilistic logic neural networks for reasoning. Advances in Neural Information Processing Systems32 (2019)

    2019

  11. [18]

    Ali Sadeghian, Mohammadreza Armandpour, Patrick Ding, and Daisy Zhe Wang

  12. [19]

    Michael Schlichtkrull, Thomas N Kipf, Peter Bloem, Rianne van den Berg, Ivan Titov, and Max Welling. 2018. Modeling relational data with graph convolutional networks. InProceedings of European Semantic Web Conference. 593–607

    2018

  13. [20]

    Zhixiang Su, Di Wang, Chunyan Miao, and Lizhen Cui. 2023. Multi-Aspect Ex- plainable Inductive Relation Prediction by Sentence Transformer. InProceedings of the AAAI Conference on Artificial Intelligence. 6533–6540

    2023

  14. [21]

    Komal Teru, Etienne Denis, and Will Hamilton. 2020. Inductive relation prediction by subgraph reasoning. InProceedings of International Conference on Machine Learning. 9448–9457

    2020

  15. [22]

    Kristina Toutanova and Danqi Chen. 2015. Observed versus latent features for knowledge base and text inference. InProceedings of the 3rd workshop on continuous vector space models and their compositionality. 57–66

    2015

  16. [23]

    Shikhar Vashishth, Soumya Sanyal, Vikram Nitin, and Partha Talukdar. 2019. Composition-based multi-relational graph convolutional networks.arXiv preprint arXiv:1911.03082(2019)

    2019 arXiv

  17. [24]

    Xiang Wang, Xiangnan He, Yixin Cao, Meng Liu, and Tat-Seng Chua. 2019. Kgat: Knowledge graph attention network for recommendation. InProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining. 950–958

    2019

  18. [25]

    Zonghan Wu, Shirui Pan, Fengwen Chen, Guodong Long, Chengqi Zhang, and S Yu Philip. 2020. A comprehensive survey on graph neural networks.IEEE transactions on neural networks and learning systems32, 1 (2020), 4–24

    2020

  19. [26]

    Wenhan Xiong, Thien Hoang, and William Yang Wang. 2017. Deeppath: A reinforcement learning method for knowledge graph reasoning.arXiv preprint arXiv:1707.06690(2017)

    2017 arXiv

  20. [27]

    Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. 2018. How powerful are graph neural networks?arXiv preprint arXiv:1810.00826(2018)

    2018 arXiv

  21. [28]

    Xiaoran Xu, Wei Feng, Yunsheng Jiang, Xiaohui Xie, Zhiqing Sun, and Zhi-Hong Deng. 2019. Dynamically pruned message passing networks for large-scale knowledge graph reasoning.arXiv preprint arXiv:1909.11334(2019)

    2019 arXiv

  22. [29]

    Fan Yang, Zhilin Yang, and William W Cohen. 2017. Differentiable learning of logical rules for knowledge base reasoning.Advances in Neural Information Processing Systems30 (2017)

    2017

  23. [30]

    Zhitao Ying, Jiaxuan You, Christopher Morris, Xiang Ren, Will Hamilton, and Jure Leskovec. 2018. Hierarchical graph representation learning with differentiable pooling.Advances in Neural Information Processing Systems31 (2018)

    2018

  24. [32]

    Zhanqiu Zhang, Jie Wang, Jieping Ye, and Feng Wu. 2022. Rethinking graph convolutional networks in knowledge graph completion. InProceedings of the ACM Web Conference. 798–807

    2022

  25. [33]

    Zhaocheng Zhu, Zuobai Zhang, Louis-Pascal Xhonneux, and Jian Tang. 2021. Neural bellman-ford networks: A general graph neural network framework for link prediction.Advances in Neural Information Processing Systems34 (2021), 29476–29490. A Proof of Theorem Theorem 2.Given Assum...

    2021

  26. [2019]

    Drum: End-to-end differentiable rule mining on knowledge graphs.Ad- vances in Neural Information Processing Systems32 (2019)

    2019

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.