pith. machine review for the scientific record. sign in

arxiv: 2605.13540 · v1 · submitted 2026-05-13 · 💻 cs.LG · cs.AI

Recognition: no theorem link

Decoupled and Divergence-Conditioned Prompt for Multi-domain Dynamic Graph Foundation Models

Haonan Yuan, Jianxin Li, Junhua Shi, Philip S. Yu, Qingyun Sun, Xingcheng Fu

Authors on Pith no claims yet

Pith reviewed 2026-05-14 20:18 UTC · model grok-4.3

classification 💻 cs.LG cs.AI
keywords dynamic graphsgraph foundation modelsmulti-domain learningprompt tuningnegative transferdecouplinggraph neural networks
0
0 comments X

The pith

DyGFM decouples semantic and temporal patterns in dynamic graphs and uses divergence-conditioned prompts to enable effective multi-domain pre-training without negative transfer.

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

Dynamic graphs from different domains show inconsistent semantic and temporal patterns that make unified pre-training difficult and often cause negative transfer in the standard pretrain-then-finetune setup. The paper introduces DyGFM, which applies a dual-branch pre-training strategy to separate transferable semantics from domain-specific dynamics. It adds a cross-domain routing mechanism that selects experts according to measured divergence, then uses a divergence-conditioned prompt generator to create lightweight, learnable graph prompts for fast downstream adaptation. A reader would care because dynamic graphs appear in many real systems and a single model that works across domains would reduce the need to train separate models for each new source.

Core claim

DyGFM is a Dynamic Graph Foundation Model over multiple domains based on decoupled and divergence-conditioned prompting. It disentangles transferable semantics from the domain-specific dynamics through a dual-branch pre-training strategy, alleviates negative transfer during domain adaptation via a cross-domain routing mechanism with divergence-aware expert selection, and enables efficient downstream fine-tuning with a divergence-conditioned prompt generator that injects lightweight learnable graph prompts tailored to semantic and temporal traits.

What carries the argument

Dual-branch pre-training for semantic-temporal decoupling together with divergence-aware expert selection in cross-domain routing and a divergence-conditioned prompt generator.

If this is right

  • DyGFM outperforms 12 state-of-the-art baselines on node classification and link prediction across continuous dynamic graph benchmarks.
  • The model achieves higher effectiveness and efficiency than prior multi-domain or single-domain approaches.
  • Unified modeling becomes feasible for dynamic graphs whose semantic and temporal traits differ across domains.
  • Lightweight divergence-conditioned prompts allow fast adaptation to new domains without full retraining.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same decoupling idea could be tested on static graphs or other sequential data where domain shifts cause negative transfer.
  • If divergence metrics prove stable, they might serve as a general diagnostic for choosing which pre-trained components to reuse in new settings.
  • Success on continuous benchmarks suggests the method could scale to streaming scenarios with evolving domains.
  • The approach implies that foundation models for graphs may need explicit mechanisms for separating universal from domain-specific structure.

Load-bearing premise

Semantic and temporal patterns can be cleanly separated and divergence metrics will reliably guide expert selection to prevent negative transfer across arbitrary domains without new biases or per-domain tuning.

What would settle it

If a single-domain model trained only on the target domain outperforms DyGFM on held-out test sets from new domain combinations, the claim that the decoupling and routing reliably avoid negative transfer would be falsified.

Figures

Figures reproduced from arXiv: 2605.13540 by Haonan Yuan, Jianxin Li, Junhua Shi, Philip S. Yu, Qingyun Sun, Xingcheng Fu.

Figure 1
Figure 1. Figure 1: Challenges of constructing a dynamic GFM. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The framework of DyGFM. (1) Dual-Branch Pre-training. DyGFM decouples transferable semantics and domain-specific temporal dynamics. (2) Cross-Domain Adaptation. Divergence-aware routing selects relevant source-domain experts according to semantic and temporal discrepancies. (3) Conditioned Fine-tuning. Divergence-conditioned graph prompts enable efficient target-domain adaptation with the frozen pre-traine… view at source ↗
Figure 3
Figure 3. Figure 3: Ablation study on node classification and transductive link prediction. [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Routing Visualizations. Reddit Reddit MOOC MOOC Wikipedia Wikipedia MOOC MOOC Reddit Reddit Wikipedia (raw) Wikipedia (finetuned) Reddit (finetuned) MOOC (finetuned) Reddit (raw) MOOC (raw) Suspended Active [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: Hyper-parameter Sensitivity Analysis. Smaller values may underemphasize semantic alignment, while larger values slightly over-regularize the semantic divergence term. For λt , the model maintains stable performance within the range of 0.1 to 0.3, but performance decreases when the value becomes overly large. This indicates that moderate temporal regularization helps preserve temporal smoothness, whereas ex… view at source ↗
read the original abstract

Dynamic graphs are ubiquitous in real-world systems, and building generalizable dynamic Graph Foundation Models has become a frontier in graph learning. However, dynamic graphs from different domains pose fundamental challenges to unified modeling, as their semantic and temporal patterns are inherently inconsistent, making the multi-domain pre-training difficult. Consequently, the widely used "pretrain-then-finetune" paradigm often suffers from severe negative knowledge transfer. To the best of our knowledge, there exists no multi-domain dynamic GFM. In this work, we propose DyGFM, a Dynamic Graph Foundation Model over multiple domains based on decoupled and divergence-conditioned prompting. To disentangle transferable semantics from the domain-specific dynamics, we introduce a dual-branch pre-training strategy with semantic-temporal decoupling. To alleviate negative transfer during domain adaptation, we further develop a cross-domain routing mechanism with divergence-aware expert selection. To enable efficient downstream fine-tuning, we design a divergence-conditioned prompt generator that injects lightweight, learnable graph prompts tailored to semantic and temporal traits. Extensive experiments on continuous dynamic graph benchmarks demonstrate that DyGFM consistently outperforms 12 state-of-the-art baselines on both node classification and link prediction tasks, achieving superior effectiveness and efficiency.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The paper proposes DyGFM, a multi-domain dynamic graph foundation model that employs a dual-branch pre-training strategy with semantic-temporal decoupling to separate transferable semantics from domain-specific dynamics, a cross-domain routing mechanism with divergence-aware expert selection to mitigate negative transfer, and a divergence-conditioned prompt generator for lightweight, tailored fine-tuning. It claims consistent outperformance over 12 state-of-the-art baselines on node classification and link prediction tasks in continuous dynamic graph benchmarks, with gains in both effectiveness and efficiency.

Significance. If the empirical gains hold under rigorous controls, the work would represent a meaningful step toward generalizable dynamic graph foundation models by directly addressing domain heterogeneity through decoupling and routing, potentially reducing reliance on per-domain retraining and offering a practical path for multi-domain pre-training in graph learning.

major comments (1)
  1. [Experiments] Experimental section: the headline claim of consistent outperformance over 12 baselines requires explicit confirmation that data splits, hyperparameter search protocols, and statistical significance tests (e.g., multiple random seeds with reported p-values) were fixed in advance rather than selected post-hoc, as this directly affects whether the reported gains are load-bearing for the central multi-domain claim.
minor comments (2)
  1. [Method] Method section: the precise formulation of the divergence threshold and its interaction with the expert selection routing should be stated with an equation or pseudocode to support reproducibility.
  2. [Preliminaries] Notation: ensure consistent use of symbols for semantic and temporal branches across the dual-branch pre-training description and the prompt generator.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive feedback and positive recommendation for minor revision. We address the single major comment below by clarifying our experimental protocols and committing to explicit documentation in the revised manuscript.

read point-by-point responses

  1. Referee: [Experiments] Experimental section: the headline claim of consistent outperformance over 12 baselines requires explicit confirmation that data splits, hyperparameter search protocols, and statistical significance tests (e.g., multiple random seeds with reported p-values) were fixed in advance rather than selected post-hoc, as this directly affects whether the reported gains are load-bearing for the central multi-domain claim.

    Authors: We agree that explicit confirmation of pre-fixed protocols is essential for the credibility of the multi-domain claims. In the original experiments, all data splits were determined in advance using temporal ordering (earliest 70% for training, next 15% for validation, latest 15% for testing) to respect the continuous dynamic nature of the graphs; no post-hoc adjustments were made. Hyperparameter search followed a fixed grid-search protocol over a predefined range on the validation set only, with the same search space applied uniformly across all baselines and domains. All reported results are means and standard deviations over five independent random seeds (with seed values fixed in advance), and we computed paired t-test p-values between DyGFM and each baseline to assess statistical significance. To address the referee's concern directly, we will add a new subsection titled 'Experimental Protocol and Reproducibility' that explicitly states these fixed procedures, lists the seed values, and includes the p-values in the main result tables. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical validation of engineering choices

full rationale

The paper introduces DyGFM as an architectural proposal relying on dual-branch pre-training for semantic-temporal decoupling, divergence-aware expert routing, and a conditioned prompt generator. These components are framed as design decisions to address negative transfer, not as quantities derived from equations that reduce to the inputs by construction. Central claims rest on empirical outperformance versus 12 baselines on node classification and link prediction tasks across continuous dynamic graph benchmarks. No load-bearing self-citations, fitted parameters renamed as predictions, or ansatzes smuggled via prior work appear in the derivation chain. The model is self-contained as an engineering contribution whose validity is assessed externally through experiments rather than internal tautology.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

The design rests on the unproven premise that semantic and temporal signals can be cleanly separated by dual branches and that divergence metrics will select experts without circular dependence on the same data used for evaluation. No machine-checked proofs or parameter-free derivations are referenced.

free parameters (2)
  • divergence threshold for expert selection
    Chosen to control routing; value not stated in abstract and must be tuned per domain set.
  • prompt generator hidden size
    Lightweight learnable parameter whose dimension is a design choice fitted during fine-tuning.
axioms (1)
  • domain assumption Semantic and temporal patterns in dynamic graphs can be disentangled without loss of critical information
    Invoked to justify the dual-branch pre-training strategy.

pith-pipeline@v0.9.0 · 5522 in / 1276 out tokens · 33372 ms · 2026-05-14T20:18:30.424750+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

122 extracted references · 98 canonical work pages · 4 internal anchors

  1. [1]

    Random graph models of social networks,

    M. E. Newman, D. J. Watts, and S. H. Strogatz, “Random graph models of social networks,”Proceedings of the National Academy of Sciences, vol. 99, no. suppl 1, pp. 2566–2572, 2002

    2002

  2. [2]

    Graph neural networks for social recommendation,

    W. Fan, Y . Ma, Q. Li, Y . He, E. Zhao, J. Tang, and D. Yin, “Graph neural networks for social recommendation,” inWWW, 2019, pp. 417–426

    2019

  3. [3]

    Graph neural networks for friend ranking in large-scale social platforms,

    A. Sankar, Y . Liu, J. Yu, and N. Shah, “Graph neural networks for friend ranking in large-scale social platforms,” inWWW, 2021, pp. 2535–2546

    2021

  4. [4]

    Graph convolutional neural networks for web-scale recommender systems,

    R. Ying, R. He, K. Chen, P. Eksombatchai, W. L. Hamilton, and J. Leskovec, “Graph convolutional neural networks for web-scale recommender systems,” inKDD, 2018, pp. 974–983

    2018

  5. [5]

    Session-based recommendation with graph neural networks,

    S. Wu, Y . Tang, Y . Zhu, L. Wang, X. Xie, and T. Tan, “Session-based recommendation with graph neural networks,”AAAI, vol. 33, no. 01, pp. 346–353, 2019

    2019

  6. [6]

    Knowledge-aware graph neural networks with label smoothness regularization for recommender systems,

    H. Wang, F. Zhang, M. Zhang, J. Leskovec, M. Zhao, W. Li, and Z. Wang, “Knowledge-aware graph neural networks with label smoothness regularization for recommender systems,” inKDD, 2019, pp. 968–977

    2019

  7. [7]

    Knowledge graph embedding by translating on hyperplanes,

    Z. Wang, J. Zhang, J. Feng, and Z. Chen, “Knowledge graph embedding by translating on hyperplanes,” inAAAI, vol. 28, no. 1, 2014

    2014

  8. [8]

    Knowledge graph embedding: A survey of approaches and applications,

    Q. Wang, Z. Mao, B. Wang, and L. Guo, “Knowledge graph embedding: A survey of approaches and applications,”IEEE TKDE, vol. 29, no. 12, pp. 2724–2743, 2017

    2017

  9. [9]

    Knowledge graph reasoning with relational digraph,

    Y . Zhang and Q. Yao, “Knowledge graph reasoning with relational digraph,” inWWW, 2022, pp. 912–924

    2022

  10. [10]

    KGNN: Knowledge graph neural network for drug-drug interaction prediction,

    X. Lin, Z. Quan, Z.-J. Wang, T. Ma, and X. Zeng, “KGNN: Knowledge graph neural network for drug-drug interaction prediction,” inIJCAI, 2021, pp. 2739–2745

    2021

  11. [11]

    Structure-based protein function prediction using graph convolutional networks,

    V . Gligorijevi´c, P. D. Renfrew, T. Kosciolek, J. K. Leman, D. Berenberg, T. Vatanen, C. Chandler, B. C. Taylor, I. M. Fisk, H. Vlamakiset al., “Structure-based protein function prediction using graph convolutional networks,”Nature Communications, vol. 12, no. 1, p. 3168, 2021

    2021

  12. [12]

    Prediction of protein-protein interaction using graph neural networks,

    K. Jha, S. Saha, and H. Singh, “Prediction of protein-protein interaction using graph neural networks,”Scientific Reports, vol. 12, no. 1, p. 8360, 2022

    work page 2022

  13. [13]

    Towards event prediction in temporal graphs,

    W. Fan, R. Jin, P. Lu, C. Tian, and R. Xu, “Towards event prediction in temporal graphs,”VLDB, vol. 15, no. 9, pp. 1861–1874, 2022

    work page 2022

  14. [14]

    TREND: Temporal event and node dynamics for graph representation learning,

    Z. Wen and Y . Fang, “TREND: Temporal event and node dynamics for graph representation learning,” inWWW, 2022, pp. 1159–1169

    work page 2022

  15. [15]

    TMac: Temporal multi-modal graph learning for acoustic event classification,

    M. Liu, K. Liang, D. Hu, H. Yu, Y . Liu, L. Meng, W. Tu, S. Zhou, and X. Liu, “TMac: Temporal multi-modal graph learning for acoustic event classification,” inACM MM, 2023, pp. 3365–3374

    work page 2023

  16. [16]

    Spatio-temporal attentive RNN for node classification in temporal attributed graphs,

    D. Xu, W. Cheng, D. Luo, X. Liu, and X. Zhang, “Spatio-temporal attentive RNN for node classification in temporal attributed graphs,” in IJCAI, 2019, pp. 3947–3953

    work page 2019

  17. [17]

    Node embedding over temporal graphs,

    U. Singer, I. Guy, and K. Radinsky, “Node embedding over temporal graphs,” inIJCAI, 2019, pp. 4605–4612

    work page 2019

  18. [18]

    Temporal graph benchmark for machine learning on temporal graphs,

    S. Huang, F. Poursafaei, J. Danovitch, M. Fey, W. Hu, E. Rossi, J. Leskovec, M. Bronstein, G. Rabusseau, and R. Rabbany, “Temporal graph benchmark for machine learning on temporal graphs,”NeurIPS, vol. 36, pp. 2056–2073, 2023

    work page 2056

  19. [19]

    Spatio-temporal graph structure learning for traffic forecasting,

    Q. Zhang, J. Chang, G. Meng, S. Xiang, and C. Pan, “Spatio-temporal graph structure learning for traffic forecasting,” inAAAI, vol. 34, no. 01, 2020, pp. 1177–1185

    work page 2020

  20. [20]

    Neural temporal walks: Motif-aware representation learning on continuous-time dynamic graphs,

    M. Jin, Y .-F. Li, and S. Pan, “Neural temporal walks: Motif-aware representation learning on continuous-time dynamic graphs,”NeurIPS, vol. 35, pp. 19 874–19 886, 2022

    work page 2022

  21. [21]

    Time-aware dynamic graph embedding for asynchronous structural evolution,

    Y . Yang, H. Yin, J. Cao, T. Chen, Q. V . H. Nguyen, X. Zhou, and L. Chen, “Time-aware dynamic graph embedding for asynchronous structural evolution,”IEEE TKDE, vol. 35, no. 9, pp. 9656–9670, 2023

    work page 2023

  22. [22]

    Continuous graph neural networks,

    L.-P. A. Xhonneux, M. Qu, and J. Tang, “Continuous graph neural networks,” inICML, 2020, pp. 10 432–10 441

    2020

  23. [23]

    Foundations and modeling of dynamic networks using dynamic graph neural networks: A survey,

    J. Skarding, B. Gabrys, and K. Musial, “Foundations and modeling of dynamic networks using dynamic graph neural networks: A survey,” IEEE Access, vol. 9, pp. 79 143–79 168, 2021

    2021

  24. [24]

    Dynamic graph neural networks for sequential recommendation,

    M. Zhang, S. Wu, X. Yu, Q. Liu, and L. Wang, “Dynamic graph neural networks for sequential recommendation,”IEEE TKDE, vol. 35, no. 5, pp. 4741–4753, 2022

    2022

  25. [25]

    Calendar graph neural networks for modeling time structures in spatiotemporal user behaviors,

    D. Wang, M. Jiang, M. Syed, O. Conway, V . Juneja, S. Subramanian, and N. V . Chawla, “Calendar graph neural networks for modeling time structures in spatiotemporal user behaviors,” inKDD, 2020, pp. 2581–2589

    2020

  26. [26]

    TP-GNN: Continuous dynamic graph neural network for graph classification,

    J. Liu, J. Liu, K. Zhao, Y . Tang, and W. Chen, “TP-GNN: Continuous dynamic graph neural network for graph classification,” inIEEE ICDE, 2024, pp. 2848–2861

    2024

  27. [27]

    Boosting GNN- based link prediction via pu-auc optimization,

    Y . Mao, Y . Hao, X. Cao, Y . Gao, C. Yao, and X. Lin, “Boosting GNN- based link prediction via pu-auc optimization,”IEEE TKDE, 2025

    2025

  28. [28]

    BRIGHT-graph neural networks in real-time fraud detection,

    M. Lu, Z. Han, S. X. Rao, Z. Zhang, Y . Zhao, Y . Shan, R. Raghunathan, C. Zhang, and J. Jiang, “BRIGHT-graph neural networks in real-time fraud detection,” inCIKM, 2022, pp. 3342–3351

    2022

  29. [29]

    DGA- GNN: Dynamic grouping aggregation GNN for fraud detection,

    M. Duan, T. Zheng, Y . Gao, G. Wang, Z. Feng, and X. Wang, “DGA- GNN: Dynamic grouping aggregation GNN for fraud detection,” in AAAI, vol. 38, no. 10, 2024, pp. 11 820–11 828

    2024

  30. [30]

    Dynamic fraud detection: Integrating reinforcement learning into graph neural networks,

    Y . Dong, J. Yao, J. Wang, Y . Liang, S. Liao, and M. Xiao, “Dynamic fraud detection: Integrating reinforcement learning into graph neural networks,” in6th International Conference on Data-driven Optimization of Complex Systems. IEEE, 2024, pp. 818–823

    work page 2024

  31. [31]

    Predicting drug-protein interaction using quasi-visual question answering system,

    S. Zheng, Y . Li, S. Chen, J. Xu, and Y . Yang, “Predicting drug-protein interaction using quasi-visual question answering system,”Nature Machine Intelligence, vol. 2, no. 2, pp. 134–140, 2020

    work page 2020

  32. [32]

    Graph neural networks accelerated molecular dynamics,

    Z. Li, K. Meidani, P. Yadav, and A. Barati Farimani, “Graph neural networks accelerated molecular dynamics,”The Journal of Chemical Physics, vol. 156, no. 14, 2022

    work page 2022

  33. [33]

    A spatial-temporal graph attention network for protein-ligand binding affinity prediction based on molecular geometry,

    G. Li, Y . Yuan, and R. Zhang, “A spatial-temporal graph attention network for protein-ligand binding affinity prediction based on molecular geometry,”Multimedia Systems, vol. 31, no. 2, p. 94, 2025

    work page 2025

  34. [34]

    Prompt learning on temporal interaction graphs,

    X. Chen, S. Zhang, Y . Xiong, X. Wu, J. Zhang, X. Sun, Y . Zhang, F. Zhao, and Y . Kang, “Prompt learning on temporal interaction graphs,” arXiv preprint arXiv:2402.06326, 2024

    work page arXiv 2024

  35. [35]

    Node-time conditional prompt learning in dynamic graphs,

    X. Yu, Z. Liu, X. Zhang, and Y . Fang, “Node-time conditional prompt learning in dynamic graphs,” inICLR, 2025

    2025

  36. [36]

    A survey on evaluation of large language models,

    Y . Chang, X. Wang, J. Wang, Y . Wu, L. Yang, K. Zhu, H. Chen, X. Yi, C. Wang, Y . Wanget al., “A survey on evaluation of large language models,”ACM TIST, vol. 15, no. 3, pp. 1–45, 2024

    work page 2024

  37. [37]

    Vision-language models for vision tasks: A survey,

    J. Zhang, J. Huang, S. Jin, and S. Lu, “Vision-language models for vision tasks: A survey,”IEEE TPAMI, 2024

    work page 2024

  38. [38]

    GPT-4 Technical Report

    J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkatet al., “GPT-4 technical report,”arXiv preprint arXiv:2303.08774, 2023

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  39. [39]

    Transformers in vision: A survey,

    S. Khan, M. Naseer, M. Hayat, S. W. Zamir, F. S. Khan, and M. Shah, “Transformers in vision: A survey,”ACM Computing Surveys, vol. 54, no. 10s, pp. 1–41, 2022

    2022

  40. [40]

    Position: Graph foundation models are already here,

    H. Mao, Z. Chen, W. Tang, J. Zhao, Y . Ma, T. Zhao, N. Shah, M. Galkin, and J. Tang, “Position: Graph foundation models are already here,” in ICML, 2024

    work page 2024

  41. [41]

    Graph foundation model,

    C. Shi, J. Chen, J. Liu, and C. Yang, “Graph foundation model,” Frontiers of Computer Science, vol. 18, no. 6, 2024

    work page 2024

  42. [42]

    ProG: A graph prompt learning benchmark,

    C. Zi, H. Zhao, X. Sun, Y . Lin, H. Cheng, and J. Li, “ProG: A graph prompt learning benchmark,”NeurIPS, vol. 37, pp. 95 406–95 437, 2024

    work page 2024

  43. [43]

    Lecture-style tutorial: Towards graph foundation models,

    C. Shi, C. Yang, Y . Fang, L. Sun, and P. S. Yu, “Lecture-style tutorial: Towards graph foundation models,” inWWW, 2024, pp. 1264–1267

    work page 2024

  44. [44]

    On the Opportunities and Risks of Foundation Models

    R. Bommasani, D. A. Hudson, E. Adeli, R. Altman, S. Arora, S. von Arx, M. S. Bernstein, J. Bohg, A. Bosselut, E. Brunskillet al., “On the opportunities and risks of foundation models,”arXiv preprint arXiv:2108.07258, 2021

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  45. [45]

    GraphGPT: Graph instruction tuning for large language models,

    J. Tang, Y . Yang, W. Wei, L. Shi, L. Su, S. Cheng, D. Yin, and C. Huang, “GraphGPT: Graph instruction tuning for large language models,” in SIGIR, 2024, pp. 491–500

    work page 2024

  46. [46]

    UniGraph: Learning a unified cross-domain foundation model for text-attributed graphs,

    Y . He, Y . Sui, X. He, and B. Hooi, “UniGraph: Learning a unified cross-domain foundation model for text-attributed graphs,” inKDD, 2025, pp. 448–459

    work page 2025

  47. [47]

    GraphFM: A scalable framework for multi-graph pretraining,

    D. Lachi, M. Azabou, V . Arora, and E. Dyer, “GraphFM: A scalable framework for multi-graph pretraining,”arXiv preprint arXiv:2407.11907, 2024

    work page arXiv 2024

  48. [48]

    Continuous-time dynamic network embeddings,

    G. H. Nguyen, J. B. Lee, R. A. Rossi, N. K. Ahmed, E. Koh, and S. Kim, “Continuous-time dynamic network embeddings,” inWWW, 2018, pp. 969–976

    work page 2018

  49. [49]

    Inductive representation learning in temporal networks via causal anonymous walks,

    Y . Wang, Y .-Y . Chang, Y . Liu, J. Leskovec, and P. Li, “Inductive representation learning in temporal networks via causal anonymous walks,” inICLR, 2021

    work page 2021

  50. [50]

    Provably expressive temporal graph networks,

    A. Souza, D. Mesquita, S. Kaski, and V . Garg, “Provably expressive temporal graph networks,”NeurIPS, vol. 35, pp. 32 257–32 269, 2022

    work page 2022

  51. [51]

    Improving temporal link prediction via temporal walk matrix projection,

    X. Lu, L. Sun, T. Zhu, and W. Lv, “Improving temporal link prediction via temporal walk matrix projection,”NeurIPS, vol. 37, pp. 141 153– 141 182, 2024

    work page 2024

  52. [52]

    Inductive representation learning on temporal graphs,

    D. Xu, C. Ruan, E. Korpeoglu, S. Kumar, and K. Achan, “Inductive representation learning on temporal graphs,” inICLR, 2020

    work page 2020

  53. [53]

    Do we really need complicated model architectures for temporal networks?

    W. Cong, S. Zhang, J. Kang, B. Yuan, H. Wu, X. Zhou, H. Tong, and M. Mahdavi, “Do we really need complicated model architectures for temporal networks?” inICLR, 2023. YUANet al.: DECOUPLED AND DIVERGENCE-CONDITIONED PROMPT FOR MULTI-DOMAIN DYNAMIC GRAPH FOUNDATION MODELS 17

    work page 2023

  54. [54]

    Towards better dynamic graph learning: New architecture and unified library,

    L. Yu, L. Sun, B. Du, and W. Lv, “Towards better dynamic graph learning: New architecture and unified library,” inNeurIPS, 2023, pp. 67 686–67 700

    work page 2023

  55. [55]

    Co-neighbor encoding schema: A light-cost structure encoding method for dynamic link prediction,

    K. Cheng, P. Linzhi, J. Ye, L. Sun, and B. Du, “Co-neighbor encoding schema: A light-cost structure encoding method for dynamic link prediction,” inKDD, 2024, pp. 421–432

    work page 2024

  56. [56]

    Towards adaptive neighborhood for advancing temporal interaction graph modeling,

    S. Zhang, X. Chen, Y . Xiong, X. Wu, Y . Zhang, Y . Fu, Y . Zhao, and J. Zhang, “Towards adaptive neighborhood for advancing temporal interaction graph modeling,” inKDD, 2024, pp. 4290–4301

    work page 2024

  57. [57]

    Repeat-aware neighbor sampling for dynamic graph learning,

    T. Zou, Y . Mao, J. Ye, and B. Du, “Repeat-aware neighbor sampling for dynamic graph learning,” inKDD, 2024, pp. 4722–4733

    work page 2024

  58. [58]

    Predicting path failure in time-evolving graphs,

    J. Li, Z. Han, H. Cheng, J. Su, P. Wang, J. Zhang, and L. Pan, “Predicting path failure in time-evolving graphs,” inKDD, 2019, pp. 1279–1289

    work page 2019

  59. [59]

    Streaming graph neural networks,

    Y . Ma, Z. Guo, Z. Ren, J. Tang, and D. Yin, “Streaming graph neural networks,” inSIGIR, 2020, pp. 719–728

    work page 2020

  60. [60]

    Recurrent temporal revision graph networks,

    Y . Chen, A. Zeng, Q. Yu, K. Zhang, C. Yuanpeng, K. Wu, G. Huzhang, H. Yu, and Z. Zhou, “Recurrent temporal revision graph networks,” NeurIPS, vol. 36, pp. 69 348–69 360, 2023

    work page 2023

  61. [61]

    Embedding temporal network via neighborhood formation,

    Y . Zuo, G. Liu, H. Lin, J. Guo, X. Hu, and J. Wu, “Embedding temporal network via neighborhood formation,” inKDD, 2018, pp. 2857–2866

    work page 2018

  62. [62]

    DyRep: Learning representations over dynamic graphs,

    R. Trivedi, M. Farajtabar, P. Biswal, and H. Zha, “DyRep: Learning representations over dynamic graphs,” inICLR, 2019

    work page 2019

  63. [63]

    Temporal network embedding with micro-and macro-dynamics,

    Y . Lu, X. Wang, C. Shi, P. S. Yu, and Y . Ye, “Temporal network embedding with micro-and macro-dynamics,” inCIKM, 2019, pp. 469– 478

    work page 2019

  64. [64]

    EasyDGL: Encode, train and interpret for continuous-time dynamic graph learning,

    C. Chen, H. Geng, N. Yang, X. Yang, and J. Yan, “EasyDGL: Encode, train and interpret for continuous-time dynamic graph learning,”TPAMI, vol. 46, no. 12, pp. 10 845–10 862, 2024

    work page 2024

  65. [65]

    Temporal graph networks for deep learning on dynamic graphs,

    E. Rossi, B. Chamberlain, F. Frasca, D. Eynard, F. Monti, and M. Bronstein, “Temporal graph networks for deep learning on dynamic graphs,”arXiv preprint arXiv:2006.10637, 2020

    work page arXiv 2006

  66. [66]

    Neighborhood-aware scalable temporal network representation learning,

    Y . Luo and P. Li, “Neighborhood-aware scalable temporal network representation learning,” inLOG. PMLR, 2022, pp. 1–1

    work page 2022

  67. [67]

    TIGER: Temporal interaction graph embedding with restarts,

    Y . Zhang, Y . Xiong, Y . Liao, Y . Sun, Y . Jin, X. Zheng, and Y . Zhu, “TIGER: Temporal interaction graph embedding with restarts,” inWWW, 2023, pp. 478–488

    work page 2023

  68. [68]

    PRES: Toward scalable memory-based dynamic graph neural networks,

    J. Su, D. Zou, and C. Wu, “PRES: Toward scalable memory-based dynamic graph neural networks,” inICLR, 2020

    work page 2020

  69. [69]

    MemMap: An adaptive and latent memory structure for dynamic graph learning,

    S. Ji, M. Liu, L. Sun, C. Liu, and T. Zhu, “MemMap: An adaptive and latent memory structure for dynamic graph learning,” inKDD, 2024, pp. 1257–1268

    work page 2024

  70. [70]

    MSPipe: Efficient temporal GNN training via staleness-aware pipeline,

    G. Sheng, J. Su, C. Huang, and C. Wu, “MSPipe: Efficient temporal GNN training via staleness-aware pipeline,” inKDD, 2024, pp. 2651– 2662

    work page 2024

  71. [71]

    FreeDyG: Frequency enhanced continuous- time dynamic graph model for link prediction,

    Y . Tian, Y . Qi, and F. Guo, “FreeDyG: Frequency enhanced continuous- time dynamic graph model for link prediction,” inICLR, 2024

    work page 2024

  72. [72]

    Ranking on dynamic graphs: An effective and robust band-pass disentangled approach,

    Y . Li, Y . Xu, X. Lin, W. Zhang, and Y . Zhang, “Ranking on dynamic graphs: An effective and robust band-pass disentangled approach,” in WWW, 2025, pp. 3918–3929

    work page 2025

  73. [73]

    Continuous- time sequential recommendation with temporal graph collaborative transformer,

    Z. Fan, Z. Liu, J. Zhang, Y . Xiong, L. Zheng, and P. S. Yu, “Continuous- time sequential recommendation with temporal graph collaborative transformer,” inCIKM, 2021, pp. 433–442

    work page 2021

  74. [74]

    Position- enhanced and time-aware graph convolutional network for sequential recommendations,

    L. Huang, Y . Ma, Y . Liu, B. Danny Du, S. Wang, and D. Li, “Position- enhanced and time-aware graph convolutional network for sequential recommendations,”ACM TOIS, vol. 41, no. 1, pp. 1–32, 2023

    work page 2023

  75. [75]

    TCGC: Temporal collaboration-aware graph co-evolution learning for dynamic recommendation,

    H. Tang, S. Wu, X. Sun, J. Zeng, G. Xu, and Q. Li, “TCGC: Temporal collaboration-aware graph co-evolution learning for dynamic recommendation,”ACM TOIS, vol. 43, no. 1, pp. 1–27, 2025

    work page 2025

  76. [76]

    Neural Kalman filtering for robust temporal recommendation,

    J. Xia, D. Li, H. Gu, T. Lu, P. Zhang, L. Shang, and N. Gu, “Neural Kalman filtering for robust temporal recommendation,” inWSDM, 2024, pp. 836–845

    work page 2024

  77. [77]

    Temporal graph contrastive learning for sequential recommendation,

    S. Zhang, L. Chen, C. Wang, S. Li, and H. Xiong, “Temporal graph contrastive learning for sequential recommendation,” inAAAI, vol. 38, no. 8, 2024, pp. 9359–9367

    work page 2024

  78. [78]

    A data-driven graph generative model for temporal interaction networks,

    D. Zhou, L. Zheng, J. Han, and J. He, “A data-driven graph generative model for temporal interaction networks,” inKDD, 2020, pp. 401–411

    work page 2020

  79. [79]

    SAD: semi-supervised anomaly detection on dynamic graphs,

    S. Tian, J. Dong, J. Li, W. Zhao, X. Xu, B. Wang, B. Song, C. Meng, T. Zhang, and L. Chen, “SAD: semi-supervised anomaly detection on dynamic graphs,” inIJCAI, 2023, pp. 2306–2314

    work page 2023

  80. [80]

    A generalizable anomaly detection method in dynamic graphs,

    X. Yang, X. Zhao, and Z. Shen, “A generalizable anomaly detection method in dynamic graphs,” inAAAI, vol. 39, no. 20, 2025, pp. 22 001– 22 009

    work page 2025

Showing first 80 references.