Pith. sign in

REVIEW 4 major objections 6 minor 74 references

When Speed meets Accuracy: an Efficient and Effective Graph Model for Temporal Link Prediction

T0 review · 4 major / 6 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 This paper claims that temporal link prediction needs no complex T-GNN: a hybrid of a recent-neighbor score and a shared-influence PageRank score matches or beats state-of-the-art models while running over 50x faster.

desk verdict EAGLE is a genuinely simple and fast temporal link predictor whose structural module is more empirically grounded than the stress-test suggests; the main soft spots are overbroad speedup claims and a hand-tuned hybrid with a fixed lambda. read the letter →

arxiv 2507.13825 v1 pith:XAZDG5Y4 submitted 2025-07-18 cs.AI

Haoyang Li , Yuming Xu , Yiming Li , Hanmo Liu , Darian Li , Chen Jason Zhang , Lei Chen , Qing Li This is my paper
classification cs.AI
keywords temporallinkpredictiongraphneuralnetworkspersonalizedPageRankdynamicgraphsrecentneighboraggregationefficientlearningrepresentation
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 tries to establish that the two attributes that actually matter for temporal link prediction are a node's most recent neighbors and the globally influential nodes it shares with a candidate partner, and that a model built from just these two signals can beat complicated temporal graph neural networks. EAGLE combines a time-aware module, which averages features of a node's $k_r$ most recent neighbors, with a structure-aware module that sums products of temporal personalized PageRank scores over shared top-$k_s$ influential nodes. An adaptive weighting term shifts between the two depending on how recently each node has interacted, so sparse or stale graphs fall back on structural evidence. On seven real-world temporal graphs the full model consistently matches or outperforms state-of-the-art T-GNNs, and the paper reports a more than 50x speedup over transformer-based T-GNNs. If true, this means the expensive machinery of multi-hop message passing, memory banks, and transformer encoders is not necessary for strong temporal link prediction.

What carries the argument

The load-bearing object is the structure-aware score in Equation (10): for a candidate pair $(v,u)$, sum $\pi_v(t)[v_i] \cdot \pi_u(t)[v_i]$ over the shared nodes in the top-$k_s$ entries of both endpoints' temporal personalized PageRank vectors. It is the paper's proxy for long-term global structural dependence, and it carries the whole 'global patterns' half of the argument. This score is training-free and deterministic, can be updated incrementally as new events arrive, and is motivated by a cited theorem that multi-hop GCN propagation concentrates on paths that PPR ranks. The time-aware module, by contrast, is the learned half: a two-layer MLP over the averaged features of each node's $k_r$ most recent neighbors. The adaptive weight in Equation (11) is what decides, per node pair, how much of the final score comes from each side.

What would settle it

Take any benchmark graph and keep the set of top-$k_s$ nodes fixed but replace their T-PPR scores in Equation (10) with random values drawn from the same range. If EAGLE-Hybrid's AP/MRR drops only a little, the structural module's contribution is not the T-PPR ranking itself; if it collapses, the specific T-PPR values are doing the work. A complementary test: on a graph with strong recency-driven links and no recurring hubs, EAGLE-Hybrid should reduce to EAGLE-Time via the adaptive weight; if the structural term still hurts, the weight mechanism is over-trusting structure.

Watch

Extended reading notes

Core claim

EAGLE's central discovery is that a hybrid of two simple scoring functions is sufficient for state-of-the-art temporal link prediction. The time-aware score represents each endpoint by averaging the concatenated node and edge features of its $k_r$ most recent neighbors, then passes the concatenation of the two endpoint representations through a two-layer MLP. The structure-aware score is computed between endpoints $v$ and $u$ as the sum, over nodes that appear in both of their top-$k_s$ temporal personalized PageRank vectors, of $\pi_v(t)[v_i]\cdot \pi_u(t)[v_i]$, with no learned parameters. The final hybrid score weights the time-aware score by $\exp(-\bar{t}_v)+\exp(-\bar{t}_u)$ and adds the structural score. Across seven datasets, the hybrid beats each module alone, and the paper reports effectiveness at least matching existing T-GNNs while training and inference become dramatically faster and lighter.

Load-bearing premise

The hybrid's effectiveness rests on the premise that two nodes are more likely to link when they share many of the same globally influential neighbors, with influence measured by temporal personalized PageRank; if that shared-influence score does not track future links in evolving graphs, the whole structural half of the model fails.

Editorial extensions

If this is right

  • Inference for a node pair touches only $k_r$ most recent neighbors and $k_s$ top T-PPR nodes, so runtime grows with $k_r + k_s$ rather than with multi-hop neighborhoods or a full transformer context.
  • The structure-aware score is training-free, so the hybrid can be deployed with almost no learned parameters beyond a two-layer MLP for the time-aware module.
  • On datasets where recent interactions dominate, the adaptive weight automatically leans on the time-aware score; where interactions are sparse or old, it leans on the structural score.
  • Because no memory banks or multi-hop message passing are needed, training time and peak GPU memory are dramatically lower, which the paper reports as over 50x speedup versus transformer-based T-GNNs.
  • The same two representations extend to node classification, where the paper reports large gains and speedups over the same baselines in the appendix.

Reading between the lines

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

  • Beyond the paper: on graphs with stable communities and long timespans, I expect the structure-aware score from Equation (10) to contribute most, while on bursty recency-driven graphs the adaptive weight should down-weight it; the seven benchmarks do not isolate this axis explicitly.
  • Beyond the paper: the same T-PPR shared-influence score could serve as a cheap candidate generator or pre-filter ahead of a heavier reranker, since it is training-free and deterministic.
  • Beyond the paper: an ablation that replaces the T-PPR values in Equation (10) with binary indicators of shared top-$k_s$ nodes would test whether it is the ranking or the magnitudes that matter; the paper does not run this variant.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 6 minor

Summary. The paper proposes EAGLE, a lightweight model for temporal link prediction. It consists of a time-aware module that averages node/edge features over each node's k_r most recent neighbors, and a structure-aware module that computes a score from the sum of products of T-PPR scores over common top-k_s influential nodes. The two scores are combined in Eq. (11) with a validation-tuned scalar lambda. Experiments on seven temporal graph benchmarks compare EAGLE against JODIE, TGAT, TGN, GraphMixer, Zebra, and DyGFormer, reporting AP/MRR/HR@10 and claiming more than a 50x speedup over transformer-based T-GNNs. The paper also includes an appendix on node classification.

Significance. If the structural heuristic in Eq. (10) is valid on temporal graphs, the paper provides a meaningful simplifying result: a two-module model with no multi-hop message passing and no memory banks can match or outperform complex T-GNNs on standard benchmarks. The paper ships an artifact, uses widely adopted benchmarks, and includes ablations and parameter sensitivity studies. However, the central structural score is not validated on temporal graphs, and the headline efficiency claim is not documented across all seven datasets. These gaps currently prevent the significance from being fully established.

major comments (4)
  1. [§3.2.2, Eq. (10)] The structure-aware score is the main engine of the reported gains on Wikipedia and Reddit: EAGLE-Time achieves AP 69.18 and 31.96 while EAGLE-Hybrid achieves 87.02 and 78.09. Yet Eq. (10) is supported only by a static-graph PPR experiment (Fig. 2) and a static GCN theorem (Thm. 3.1). Neither support establishes that, on a temporal graph, two nodes sharing top-k_s T-PPR nodes are more likely to interact, nor that the product form is appropriate. Please provide a temporal validation of Eq. (10): for example, compare EAGLE-Struc against static-PPR and common-neighbor baselines on the temporal benchmarks, and report the predictive power of the structural term alone with negative controls. Without this, the 'consistently superior' claim rests on an unvalidated heuristic.
  2. [§3.2.3, Eq. (11)] The hybrid score adds an unbounded structural term to a bounded sigmoid time score, and the only adaptation is a validation-tuned scalar lambda multiplied by exp(-tbar_v)+exp(-tbar_u). This mechanism cannot downweight the structural term when it is uninformative, and the structural term itself has no recency adaptivity. Please report the selected lambda values per dataset and provide ablation results with normalized structural scores (e.g., rank-based or min-max scaling) to support the claim that the model 'balances' the two signals.
  3. [§4.2.2, Table 4] The abstract's claim of 'more than 50x speedup over effective transformer-based T-GNNs' is not substantiated. Table 4 reports efficiency for only three of the seven datasets (Contacts, Wikipedia, WikiTalk), and DyGFormer either OOMs or times out on two of them; the only completed comparison is Wikipedia, where the speedup is about 91x. Please report efficiency for all datasets on which baselines complete, and state the exact comparison underlying the 50x claim. In addition, Algorithm 1 delegates the T-PPR update to the black-box procedure T-PPR_UPDATING from [25]; because the stated O(n k_s log k_s) update and O(k_s |V|) memory are imported from that reference, the paper should specify or implement the update, or explicitly bound the claims by the cited results.
  4. [§4.2.1, Table 3] Several 'superior' differences are within one standard deviation of the best baseline: for example, Wikipedia AP is 87.02±0.29 for EAGLE-Hybrid versus 86.96±0.22 for DyGFormer, and Reddit AP is 78.09±0.66 versus 77.92±0.33. To support 'consistently superior performance', please report the number of seeds and run statistical significance tests or provide a clear statement of which differences are significant.
minor comments (6)
  1. [§3.2.2] The case where the intersection of the top-k_s T-PPR sets is empty is not discussed; if the structural score is zero for many test pairs, ranking behavior should be analyzed. Please report the fraction of test pairs with a non-empty intersection.
  2. [Appendix A.1, Eq. (14)] The concatenation [h^r_v(t), h^r_v(t)] appears to be a typo for [h^r_v(t), h^s_v(t)], since the structure-aware representation is otherwise unused in the classification head.
  3. [§4.2.2, Table 4] Efficiency is omitted for AskUbuntu, SuperUser, LastFM, and Reddit; please add these results or explain why they are not reported.
  4. [§3.2.4] The statement that EAGLE-Struc is 'training-free' is misleading because alpha, beta, and k_s are tuned on validation data; clarify that 'training-free' means no gradient-based optimization.
  5. [§3.1.1, Figure 1] The motivational experiment does not state which base model is used; please specify the predictor and the number of repeats so the observation is reproducible.
  6. [§3.2.3] The phrase 'adaptive weighting mechanism' is stronger than what is implemented: lambda is a single scalar tuned on the validation set, so the mechanism is static after validation. Consider describing it as a validation-tuned trade-off parameter.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular derivation: EAGLE's claims are supported by held-out test evaluation; the only self-citation (Zebra T-PPR) is a non-load-bearing component.

full rationale

No circular derivation is present. EAGLE's time-aware score (Eq. 7) is an MLP trained on labeled historical interactions and evaluated on held-out test edges with 99 negative samples per positive; hyperparameters (α, β, k_r, k_s, λ) are selected by validation-set performance (Sec. 4.1.4), not fit to the test target. The structure-aware score (Eq. 10) is a fixed, training-free heuristic computed from T-PPR scores; although its temporal support rests on the authors' earlier Zebra paper [25], that self-citation supplies an algorithmic component and its update-complexity bound, and EAGLE is benchmarked independently against Zebra as a baseline rather than deriving its accuracy from Zebra's results. Theorem 3.1 is cited from external work (Xu et al. [56]) and is used only as motivation. The hybrid score (Eq. 11) is a weighted sum of two independently computed scores with a validation-tuned scalar λ; it is not defined in terms of the prediction target. The claimed gains and speedups are empirical numbers from Tables 3-4 and 6-7. The main risk is correctness, not circularity: Eq. (10)'s predictive validity on temporal graphs is supported only by static-graph experiments and a static GCN theorem, so if shared top-k_s T-PPR nodes do not signal future links, the hybrid model degrades toward EAGLE-Time. That is an unvalidated assumption, not a reduction of the output to the input.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The model depends on standard neural training plus two heuristic selection mechanisms. No new physical entities are introduced; the T-PPR machinery is inherited from Zebra [25], with hyperparameters tuned on validation.

free parameters (6)
  • lambda (trade-off weight) = tuned on validation set
    Eq (11) scales the temporal score by exp(-t_bar) terms; the paper calls this adaptive, but it is a single scalar chosen on the validation set.
  • alpha (T-PPR termination probability) = grid {0.1,...,1}
    Eq (8); termination probability of T-PPR; grid searched on validation.
  • beta (temporal decay factor) = grid {0.1,...,1}
    Eq (9); exponential recency decay in the transition matrix; grid searched on validation.
  • k_r (recent neighbor count) = grid {10,20,30,40,50}
    Section 3.2.1; number of most recent neighbors used by the time-aware module; tuned on validation.
  • k_s (top T-PPR node count) = grid {10,20,30,40,50}
    Section 3.2.2; number of top T-PPR nodes used by the structure-aware module; tuned on validation.
  • Time-aware MLP weights = learned by training
    Standard trained parameters W1,b1,W2,b2 in Eq (7).
assumptions (4)
  • standard math Theorem 3.1 from [56]: normalized influence in GCN-mean equals sum of path probabilities; motivates PPR as a structural importance measure.
    Invoked in Section 3.1.2 to justify using PPR for structural importance; accepted from cited literature.
  • domain assumption The T-PPR updating algorithm [25] correctly and incrementally maintains the top-k_s transition matrix on dynamic graphs.
    Algorithm 1 and the time complexity claim rely on T-PPR_UPDATING from [25]; no derivation is included here.
  • ad hoc to paper Recent neighbors and top T-PPR neighbors are sufficient statistics for temporal link prediction.
    Motivational experiments in Section 3.1 suggest this, but no formal guarantee is given; the EAGLE architecture assumes it.
  • domain assumption Node features x_{v_i}(t) of neighbors are available at current time t when computing h_v(t).
    Equation (6) uses features of neighboring nodes at time t, which in streaming settings may require feature history or static features.

how reviews work

0 comments
Cite this review

Pith. "Pith review of When Speed meets Accuracy: an Efficient and Effective Graph Model for Temporal Link Prediction." pith.science (2026). https://pith.science/paper/XAZDG5Y4

@misc{pith2026250713825,
  author       = {Pith},
  title        = {Pith review of: When Speed meets Accuracy: an Efficient and Effective Graph Model for Temporal Link Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XAZDG5Y4}},
  note         = {Machine review of arXiv:2507.13825}
}
read the original abstract

Temporal link prediction in dynamic graphs is a critical task with applications in diverse domains such as social networks, recommendation systems, and e-commerce platforms. While existing Temporal Graph Neural Networks (T-GNNs) have achieved notable success by leveraging complex architectures to model temporal and structural dependencies, they often suffer from scalability and efficiency challenges due to high computational overhead. In this paper, we propose EAGLE, a lightweight framework that integrates short-term temporal recency and long-term global structural patterns. EAGLE consists of a time-aware module that aggregates information from a node's most recent neighbors to reflect its immediate preferences, and a structure-aware module that leverages temporal personalized PageRank to capture the influence of globally important nodes. To balance these attributes, EAGLE employs an adaptive weighting mechanism to dynamically adjust their contributions based on data characteristics. Also, EAGLE eliminates the need for complex multi-hop message passing or memory-intensive mechanisms, enabling significant improvements in efficiency. Extensive experiments on seven real-world temporal graphs demonstrate that EAGLE consistently achieves superior performance against state-of-the-art T-GNNs in both effectiveness and efficiency, delivering more than a 50x speedup over effective transformer-based T-GNNs.

Figures

Figures reproduced from arXiv: 2507.13825 by the authors.

Figure 1
Figure 1. Time influence of neighbors. 1 20 40 60 80 100 ks 0.0 0.3 0.6 0.9 Cora-AP Cora-PPR Citeseer-AP Citeseer-PPR (a) Cora and Citeseer graphs. 1 20 40 60 80 100 ks 0.0 0.3 0.6 0.9 Phy-AP Phy-PPR Comp-AP Comp-PPR (b) Computer and Physics graphs [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Global influential nodes. have proven effective for improving efficiency, as demonstrated by CacheGNN [22] and Orca [24, 26], which reuse frequently accessed embeddings and intermediate results in memory, thus avoiding redundant computations during neighbor aggregation and mes￾sage passing. Unlike previous data management techniques that accelerate T-GNNs using specialized sampling, caching, or parallel processing, … view at source ↗
Figure 3
Figure 3. EAGLE framework overview. EAGLE predicts [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 6
Figure 6. Figure 6: Hit Ratio@𝑁. demonstrate that EAGLE consistently achieves higher HR@𝑁 val￾ues compared to the baselines, highlighting its superior ability to prioritize true positive nodes in its predictions. EAGLE can bal￾ance short-term temporal recency and long-term global structur…
Figure 5
Figure 5. Figure 5: Global influential node number 𝑘𝑠 . 4.4 Parameter sensitivity 4.4.1 Most Recent Neighbor Number 𝑘𝑟. We analyze the impact of the number of most recent neighbors 𝑘𝑟 in Section 3.2.1. We evaluate 𝑘𝑟 on two representative datasets, i.e., Wikipedia and AskUbuntu, vary 𝑘𝑟 ∈…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

74 extracted references · 50 canonical work pages

  1. [25]

    Yiming Li, Yanyan Shen, Lei Chen, and Mingxuan Yuan. 2023. Zebra: When tem- poral graph neural networks meet temporal personalized PageRank.Proceedings of the VLDB Endowment 16, 6 (2023), 1332–1345

    work page 2023

  2. [1]

    Thierry Bertin-Mahieux, Daniel PW Ellis, Brian Whitman, and Paul Lamere

  3. [2]

    Xin-Wei Cai, Xiangyu Ke, Kai Wang, Lu Chen, Tianming Zhang, Qing Liu, and Yunjun Gao. 2023. Efficient Temporal Butterfly Counting and Enumeration on Temporal Bipartite Graphs. Proc. VLDB Endow. 17, 4 (2023), 657–670. https: //doi.org/10.14778/3636218.3636223

    work page arXiv 2023

  4. [3]

    Chaoyi Chen, Dechao Gao, Yanfeng Zhang, Qiange Wang, Zhenbo Fu, Xuecang Zhang, Junhua Zhu, Yu Gu, and Ge Yu. 2023. NeutronStream: A Dynamic GNN Training Framework with Sliding Window for Graph Streams. Proc. VLDB Endow. 17, 3 (Nov. 2023), 455–468. https://doi.org/10.14778/3632093.3632108

    arXiv 2023

  5. [4]

    Huiyuan Chen and Jing Li. 2018. Exploiting structural and temporal evolu- tion in dynamic link prediction. In Proceedings of the 27th ACM International conference on information and knowledge management. 427–436

    work page 2018

  6. [5]

    Xin Chen, Jieming Shi, You Peng, Wenqing Lin, Sibo Wang, and Wenjie Zhang

    work page

  7. [6]

    Weilin Cong, Si Zhang, Jian Kang, Baichuan Yuan, Hao Wu, Xin Zhou, Hanghang Tong, and Mehrdad Mahdavi. 2023. Do We Really Need Complicated Model Ar- chitectures For Temporal Networks?. In The Eleventh International Conference on Learning Representations

    work page 2023

  8. [7]

    Ariel Debrouvier, Eliseo Parodi, Matías Perazzo, Valeria Soliani, and Alejandro Vaisman. 2021. A model and query language for temporal graph databases. The VLDB Journal 30, 5 (2021), 825–858

    work page 2021

Show all 74 references
  1. [8]

    Wenfei Fan, Ruochun Jin, Ping Lu, Chao Tian, and Ruiqi Xu. 2022. Towards event prediction in temporal graphs. Proceedings of the VLDB Endowment 15, 9 (2022), 1861–1874

    2022

  2. [9]

    Chen Gao, Yu Zheng, Nian Li, Yinfeng Li, Yingrong Qin, Jinghua Piao, Yuhan Quan, Jianxin Chang, Depeng Jin, Xiangnan He, et al. 2023. A survey of graph neural networks for recommender systems: Challenges, methods, and directions. ACM Transactions on Recommender Systems 1, 1 (2...

    2023

  3. [10]

    Shihong Gao, Yiming Li, Yanyan Shen, Yingxia Shao, and Lei Chen. 2024. ETC: Efficient Training of Temporal Graph Neural Networks over Large-scale Dynamic Graphs. Proceedings of the VLDB Endowment 17, 5 (2024), 1060–1072

    2024

  4. [11]

    Shihong Gao, Yiming Li, Xin Zhang, Yanyan Shen, Yingxia Shao, and Lei Chen

  5. [12]

    Johannes Gasteiger, Aleksandar Bojchevski, and Stephan Günnemann. 2018. Predict then propagate: Graph neural networks meet personalized pagerank. arXiv preprint arXiv:1810.05997 (2018)

    2018 arXiv

  6. [13]

    Julia Gastinger, Shenyang Huang, Mikhail Galkin, Erfan Loghmani, Ali Parviz, Farimah Poursafaei, Jacob Danovitch, Emanuele Rossi, Ioannis Koutis, Heiner Stuckenschmidt, Reihaneh Rabbany, and Guillaume Rabusseau. 2024. TGB 2.0: A Benchmark for Learning on Temporal Knowledge Gra...

    2024

  7. [14]

    Proceedings of the ACM on Management of Data 2, 3 (2024), 1–25

    SIMPLE: Efficient Temporal Graph Neural Network Training at Scale with Dynamic Data Placement. Proceedings of the ACM on Management of Data 2, 3 (2024), 1–25

    2024

  8. [15]

    Balázs Hidasi and Domonkos Tikk. 2012. Fast ALS-based tensor factorization for context-aware recommendation from implicit feedback. In Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2012, Bristol, UK, September 24-28, 2012. Proceedings, P...

    2012

  9. [16]

    Haixing Huang, Jinghe Song, Xuelian Lin, Shuai Ma, and Jinpeng Huai. 2016. Tgraph: A temporal graph data management system. In Proceedings of the 25th ACM International on Conference on Information and Knowledge Management. 2469–2472

    2016

  10. [17]

    Aditya Grover and Jure Leskovec. 2016. node2vec: Scalable feature learning for networks. In Proceedings of the 22nd ACM SIGKDD international conference on Knowledge discovery and data mining. 855–864

    2016

  11. [18]

    Qiang Huang, Xiao Yan, Xin Wang, Susie Xi Rao, Zhichao Han, Fangcheng Fu, Wentao Zhang, and Jiawei Jiang. 2024. Retrofitting Temporal Graph Neural Networks with Transformer. arXiv preprint arXiv:2409.05477 (2024)

    2024 arXiv

  12. [19]

    Shenyang Huang, Farimah Poursafaei, Jacob Danovitch, Matthias Fey, Weihua Hu, Emanuele Rossi, Jure Leskovec, Michael Bronstein, Guillaume Rabusseau, and Reihaneh Rabbany. 2024. Temporal graph benchmark for machine learning on temporal graphs. Advances in Neural Information Pro...

    2024

  13. [20]

    Qiang Huang, Xin Wang, Susie Xi Rao, Zhichao Han, Zitao Zhang, Yongjun He, Quanqing Xu, Yang Zhao, Zhigao Zheng, and Jiawei Jiang. 2024. Benchtemp: A General Benchmark for Evaluating Temporal Graph Neural Networks. In 2024 IEEE 40th International Conference on Data Engineering...

    2024

  14. [21]

    Jure Leskovec, Daniel Huttenlocher, and Jon Kleinberg. 2010. Governance in social media: A case study of the Wikipedia promotion process. In Proceedings of the International AAAI Conference on Web and Social Media, Vol. 4. 98–105

    2010

  15. [22]

    Haoyang Li and Lei Chen. 2021. Cache-based gnn system for dynamic graphs. In Proceedings of the 30th ACM International Conference on Information & Knowledge Management. 937–946

    2021

  16. [23]

    Srijan Kumar, Xikun Zhang, and Jure Leskovec. 2019. Predicting dynamic embedding trajectory in temporal interaction networks. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. ACM, 1269–1278

    2019

  17. [24]

    Yiming Li, Yanyan Shen, Lei Chen, and Mingxuan Yuan. 2023. Orca: Scalable tem- poral graph neural network training with theoretical guarantees. Proceedings of the ACM on Management of Data 1, 1 (2023), 1–27

    2023

  18. [26]

    Haoyang Li and Lei Chen. 2023. Early: Efficient and reliable graph neural network for dynamic graphs. Proceedings of the ACM on Management of Data 1, 2 (2023), 1–28

    2023

  19. [27]

    Zhao Li, Xin Shen, Yuhang Jiao, Xuming Pan, Pengcheng Zou, Xianling Meng, Chengwei Yao, and Jiajun Bu. 2020. Hierarchical bipartite graph neural networks: Towards large-scale e-commerce applications. In 2020 IEEE 36th International Conference on Data Engineering (ICDE). IEEE, ...

    2020

  20. [28]

    Zhao Li, Haishuai Wang, Peng Zhang, Pengrui Hui, Jiaming Huang, Jian Liao, Ji Zhang, and Jiajun Bu. 2021. Live-streaming fraud detection: A heteroge- neous graph neural network approach. InProceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. 3670–3678

    2021

  21. [29]

    Yiming Li, Yanyan Shen, Lei Chen, and Mingxuan Yuan. 2024. A Caching- based Framework for Scalable Temporal Graph Neural Network Training. ACM Transactions on Database Systems (2024)

    2024

  22. [30]

    Antonio Longa, Veronica Lachi, Gabriele Santin, Monica Bianchini, Bruno Lepri, Pietro Lió, Franco Scarselli, and Andrea Passerini. 2023. Graph Neural Networks for temporal graphs: State of the art, open challenges, and opportunities. ArXiv abs/2302.01018 (2023). https://api.se...

    2023 arXiv

  23. [31]

    Yunkai Lou, Chaokun Wang, Tiankai Gu, Hao Feng, Jun Chen, and Jeffrey Xu Yu. 2023. Time-topology analysis on temporal graphs. The VLDB Journal 32, 4 (2023), 815–843

    2023

  24. [32]

    Lipton, John Berkowitz, and Charles Elkan

    Zachary C. Lipton, John Berkowitz, and Charles Elkan. 2015. A Critical Review of Recurrent Neural Networks for Sequence Learning. arXiv:1506.00019 [cs.LG] https://arxiv.org/abs/1506.00019

    2015 arXiv

  25. [33]

    Graham K MacDonald, Kate A Brauman, Shipeng Sun, Kimberly M Carlson, Emily S Cassidy, James S Gerber, and Paul C West. 2015. Rethinking agricultural trade relationships in an era of globalization. BioScience 65, 3 (2015), 275–289

    2015

  26. [34]

    Ibomoiye Domor Mienye, Theo G Swart, and George Obaido. 2024. Recurrent neural networks: A comprehensive review of architectures, variants, and appli- cations. Information 15, 9 (2024), 517

    2024

  27. [35]

    Yuanfu Lu, Xiao Wang, Chuan Shi, Philip S Yu, and Yanfang Ye. 2019. Temporal network embedding with micro-and macro-dynamics. InProceedings of the 28th ACM international conference on information and knowledge management. 469–478

    2019

  28. [36]

    Giang Hoang Nguyen, John Boaz Lee, Ryan A Rossi, Nesreen K Ahmed, Eunyee Koh, and Sungchul Kim. 2018. Continuous-time dynamic network embeddings. In Companion proceedings of the the web conference 2018. 969–976

    2018

  29. [37]

    Ashwin Paranjape, Austin R Benson, and Jure Leskovec. 2017. Motifs in temporal networks. In Proceedings of the tenth ACM international conference on web search and data mining. 601–610

    2017

  30. [38]

    Amirhossein Nadiri and Frank W Takes. 2022. A large-scale temporal analysis of user lifespan durability on the Reddit social media platform. In Companion Proceedings of the Web Conference 2022. 677–685

    2022

  31. [39]

    Bryan Perozzi, Rami Al-Rfou, and Steven Skiena. 2014. Deepwalk: Online learning of social representations. InProceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining. 701–710

    2014

  32. [40]

    Farimah Poursafaei, Shenyang Huang, Kellin Pelrine, and Reihaneh Rabbany

  33. [41]

    Aldo Pareja, Giacomo Domeniconi, Jie Chen, Tengfei Ma, Toyotaro Suzumura, Hiroki Kanezashi, Tim Kaler, Tao Schardl, and Charles Leiserson. 2020. Evolvegcn: Evolving graph convolutional networks for dynamic graphs. In Proceedings of the AAAI conference on artificial intelligenc...

    2020

  34. [42]

    Emanuele Rossi, Ben Chamberlain, Fabrizio Frasca, Davide Eynard, Federico Monti, and Michael Bronstein. 2020. Temporal graph networks for deep learning on dynamic graphs. arXiv preprint arXiv:2006.10637 (2020)

    2020 arXiv

  35. [43]

    Kiarash Shamsi, Friedhelm Victor, Murat Kantarcioglu, Yulia Gel, and Cuneyt G Akcora. 2022. Chartalist: Labeled graph datasets for utxo and account-based blockchains. Advances in Neural Information Processing Systems 35 (2022), 34926–34939

    2022

  36. [44]

    Aneesh Sharma, Jerry Jiang, Praveen Bommannavar, Brian Larson, and Jimmy Lin. 2016. GraphJet: Real-time content recommendations at Twitter.Proceedings of the VLDB Endowment 9, 13 (2016), 1281–1292

    2016

  37. [45]

    Hongchao Qin, Rong-Hua Li, Ye Yuan, Guoren Wang, Lu Qin, and Zhiwei Zhang

  38. [46]

    Proceedings of the VLDB Endowment (2022)

    Mining bursting core in large temporal graphs. Proceedings of the VLDB Endowment (2022)

    2022

  39. [47]

    Rakshit Trivedi, Mehrdad Farajtabar, Prasenjeet Biswal, and Hongyuan Zha

  40. [48]

    A Vaswani. 2017. Attention is all you need. Advances in Neural Information Processing Systems (2017)

    2017

  41. [49]

    Gomez, Lukasz Kaiser, and Illia Polosukhin

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2023. Attention Is All You Need. arXiv:1706.03762 [cs.CL] https://arxiv.org/abs/1706.03762

    2023 arXiv

  42. [50]

    Oleksandr Shchur, Maximilian Mumme, Aleksandar Bojchevski, and Stephan Günnemann. 2018. Pitfalls of graph neural network evaluation. arXiv preprint arXiv:1811.05868 (2018)

    2018 arXiv

  43. [51]

    Jian Tang, Meng Qu, Mingzhe Wang, Ming Zhang, Jun Yan, and Qiaozhu Mei

  44. [52]

    Xuhong Wang, Ding Lyu, Mengjian Li, Yang Xia, Qi Yang, Xinwen Wang, Xin- guang Wang, Ping Cui, Yupu Yang, Bowen Sun, et al. 2021. Apan: Asynchronous propagation attention network for real-time temporal graph embedding. In Proceedings of the 2021 international conference on man...

    2021

  45. [53]

    Yanbang Wang, Yen-Yu Chang, Yunyu Liu, Jure Leskovec, and Pan Li. 2022. Inductive Representation Learning in Temporal Networks via Causal Anonymous Walks. arXiv:2101.05974 [cs.LG] https://arxiv.org/abs/2101.05974

    2022 arXiv

  46. [54]

    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 systems 32, 1 (2020), 4–24

    2020

  47. [55]

    Da Xu, Chuanwei Ruan, Evren Korpeoglu, Sushant Kumar, and Kannan Achan

  48. [56]

    Keyulu Xu, Chengtao Li, Yonglong Tian, Tomohiro Sonobe, Ken-ichi Kawarabayashi, and Stefanie Jegelka. 2018. Representation learning on graphs with jumping knowledge networks. In International conference on machine learning. PMLR, 5453–5462

    2018

  49. [57]

    Lu Wang, Xiaofu Chang, Shuang Li, Yunfei Chu, Hui Li, Wei Zhang, Xiaofeng He, Le Song, Jingren Zhou, and Hongxia Yang. 2021. Tcl: Transformer-based dy- namic graph modelling via contrastive learning.arXiv preprint arXiv:2105.07944 (2021)

    2021 arXiv

  50. [58]

    Peng Wang, BaoWen Xu, YuRong Wu, and XiaoYu Zhou. 2014. Link prediction in social networks: the state-of-the-art. arXiv preprint arXiv:1411.5118 (2014)

    2014 arXiv

  51. [59]

    Le Yu, Leilei Sun, Bowen Du, and Weifeng Lv. 2023. Towards better dynamic graph learning: New architecture and unified library. Advances in Neural Information Processing Systems 36 (2023), 67686–67700

    2023

  52. [60]

    Song Yu, Shufeng Gong, Qian Tao, Sijie Shen, Yanfeng Zhang, Wenyuan Yu, Pengxi Liu, Zhixin Zhang, Hongfu Li, Xiaojian Luo, et al . 2024. LSMGraph: A High-Performance Dynamic Graph Storage System with Multi-Level CSR. Proceedings of the ACM on Management of Data 2, 6 (2024), 1–28

    2024

  53. [61]

    Zihao Yu, Ningyi Liao, and Siqiang Luo. 2024. GENTI: GPU-powered Walk-based Subgraph Extraction for Scalable Representation Learning on Dynamic Graphs. Proceedings of the VLDB Endowment 17, 9 (2024), 2269–2278

    2024

  54. [62]

    Yalong Zhang, Rong-Hua Li, Qi Zhang, Hongchao Qin, Lu Qin, and Guoren Wang. 2024. Efficient Algorithms for Pseudoarboricity Computation in Large Static and Dynamic Graphs. Proceedings of the VLDB Endowment 17, 11 (2024), 2722–2734

    2024

  55. [63]

    Yanping Zheng, Zhewei Wei, and Jiajun Liu. 2023. Decoupled Graph Neural Networks for Large Dynamic Graphs.Proc. VLDB Endow.16, 9 (2023), 2239–2247. https://doi.org/10.14778/3598581.3598595

    2023

  56. [64]

    Yanping Zheng, Lu Yi, and Zhewei Wei. 2025. A survey of dynamic graph neural networks. Frontiers of Computer Science 19, 6 (2025), 1–18

    2025

  57. [65]

    Zhilin Yang, William Cohen, and Ruslan Salakhudinov. 2016. Revisiting semi- supervised learning with graph embeddings. In International conference on machine learning. PMLR, 40–48

    2016

  58. [66]

    Lu Yi, Jie Peng, Yanping Zheng, Fengran Mo, Zhewei Wei, Yuhang Ye, Yue Zixuan, and Zengfeng Huang. 2025. TGB-Seq Benchmark: Challenging Temporal GNNs with Complex Sequential Dynamics. InThe Thirteenth International Conference on Learning Representations. https://openreview.net...

    2025

  59. [73]

    Hongkuan Zhou, Da Zheng, Israt Nisa, Vasileios Ioannidis, Xiang Song, and George Karypis. 2022. Tgl: A general framework for temporal gnn training on billion-scale graphs. arXiv preprint arXiv:2203.14883 (2022)

    2022 arXiv

  60. [74]

    Yuan Zuo, Guannan Liu, Hao Lin, Jia Guo, Xiaoqian Hu, and Junjie Wu. 2018. Embedding temporal network via neighborhood formation. In Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining. 2857–2866. A APPENDIX A.1 Node Classification ...

    2018

  61. [2011]

    In Ismir, Vol

    The million song dataset.. In Ismir, Vol. 2. 10

  62. [2015]

    In Proceedings of the 24th international conference on world wide web

    Line: Large-scale information network embedding. In Proceedings of the 24th international conference on world wide web. 1067–1077

  63. [2019]

    In International conference on learning representations

    Dyrep: Learning representations over dynamic graphs. In International conference on learning representations

  64. [2020]

    arXiv preprint arXiv:2002.07962 (2020)

    Inductive representation learning on temporal graphs. arXiv preprint arXiv:2002.07962 (2020)

    2020 arXiv

  65. [2022]

    Towards better evaluation for dynamic link prediction.Advances in Neural Information Processing Systems 35 (2022), 32928–32941

    2022

  66. [2024]

    Proceedings of the VLDB Endowment 17, 6 (2024), 1324–1336

    Minimum Strongly Connected Subgraph Collection in Dynamic Graphs. Proceedings of the VLDB Endowment 17, 6 (2024), 1324–1336

    2024

Pith tools

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