Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

A Survey of Generalization of Graph Anomaly Detection: From Transfer Learning to Foundation Models

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper claims that all generalized graph anomaly detection methods fit under two paradigms, transfer learning and foundation models, and it supplies a systematic taxonomy to organize them.

desk verdict A useful first taxonomy for generalization in GAD, but the comprehensiveness claim outruns the documented selection process. read the letter →

arxiv 2509.06609 v1 pith:6S4LSO2D submitted 2025-09-08 cs.LG

Junjun Pan , Yu Zheng , Yue Tan , Yixin Liu This is my paper · ORCID
classification cs.LG
keywords graphanomalydetectiontransferlearningfoundationmodelsgeneralizationtaxonomyneuralnetworkszero-shotcross-domain
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

This survey sets out to establish that the scattered work on generalization in graph anomaly detection (GAD) can be organized into a single taxonomy with two main branches. The first branch, transfer learning for GAD, reuses knowledge from related source graphs to detect anomalies on a target graph. The second branch, GAD foundation models, aims for a 'one-for-all' model that scores anomalies on unseen graphs, possibly across different granularities such as nodes, edges, subgraphs, and whole graphs. The paper formalizes the problem settings for conventional GAD, transfer learning, and foundation models, and it classifies existing methods into four finer categories. A sympathetic reader would care because a reliable map lets practitioners choose methods by the kind of shift they face and lets researchers see where the design space is still empty.

What carries the argument

The load-bearing object is the taxonomy itself, built on three formal definitions: conventional GAD (training and testing on one dataset), transfer learning for GAD (using source datasets to improve detection on a target dataset), and GAD foundation models (one model scoring arbitrary samples from unseen datasets). The taxonomy's first branch is split by two challenges—learning transferable knowledge versus capturing target-specific patterns—and the second by two generalization targets—cross-granularity versus cross-scenario. These definitions do the work of the survey: they supply the coordinates in which each reviewed method is placed, and they turn the otherwise anecdotal progress of the field into a map with named design choices such as generalization-centric training, source-target representation alignment, target-aware pre-training, test-time fine-tuning, graph feature standardization, task-agnostic detection, and resource-efficient adaptation.

What would settle it

Checking every generalized GAD paper published in the five years before the survey, one could ask whether each falls cleanly into exactly one of the four taxonomy leaves; finding a peer-reviewed method that fits none of them—or that the survey's own categories force into two places at once—would falsify the claim that the taxonomy is comprehensive and systematic.

Watch

Extended reading notes

Core claim

The paper's central claim is that generalization in GAD has followed a clear evolution, and that every method along this path can be placed under one of two paradigms. In transfer learning for GAD, the model uses source-domain data to help a target domain under the assumptions that the domains share common knowledge and that their discrepancy is moderate; methods must both learn transferable knowledge and capture target-specific patterns. In GAD foundation models, a single scoring function is trained to accept arbitrary samples from unseen datasets, even from domains and granularities never seen in training; current work splits into cross-granularity models that detect anomalies at several levels at once and cross-scenario models that standardize features, learn task-agnostic detectors, and adapt efficiently. By organizing the literature this way, the authors claim to reveal the design principles behind recent generalized GAD methods and to identify open problems, including the lack of theoretical guarantees on transferability, missing standardized evaluation protocols, the road to universal one-for-all models, and human-in-the-loop adaptation.

Load-bearing premise

The survey's map is trustworthy only if the set of papers it reviews is representative of the field, but the paper gives no search strategy or inclusion criteria to guarantee that no significant generalized-GAD method is missing.

Editorial extensions

If this is right

  • A practitioner facing scarce target data can choose a transfer learning method by whether target graphs are available during training: generalization-centric training when they are not, representation alignment when they are.
  • Methods that unify node-, edge-, and graph-level detection into one framework (cross-granularity foundation models) imply that correlations across granularities can be exploited rather than treated as separate tasks.
  • The existence of zero-shot GAD foundation models implies that anomaly detection can be applied to a brand-new graph without retraining, provided features can be standardized into a common space.
  • The survey's open-challenge list implies that the field's next bottleneck is not more architectures but theory for why transfer works, unified benchmarks, and scalable one-for-all models.

Reading between the lines

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

  • The taxonomy suggests a natural test for the field: if cross-scenario foundation models keep improving, transfer learning for GAD may become a special case—when the target is known and labeled—rather than a separate paradigm.
  • Because the survey separates generalization-centric training from test-time fine-tuning, one can read future papers as combinations of these choices; a likely next step is a benchmark that compares methods across all four branches on identical shifts.
  • The emphasis on feature-space standardization implies that progress on unified graph representations, rather than detector architectures, may be what unlocks broader zero-shot generalization.
  • A reader should watch whether the categories remain stable as new methods appear: the boundary between learning transferable knowledge and capturing target-specific patterns is defined by target-data availability, which future methods may blur with continual learning.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. This paper surveys generalization in graph anomaly detection (GAD), organizing the literature into two paradigms: transfer learning and foundation models. It formalizes conventional GAD, transfer learning, and foundation-model problem settings (Definitions 1-3), proposes a two-level taxonomy (generalization-centric training vs. source-target representation alignment; target-aware pre-training vs. test-time fine-tuning; cross-granularity vs. cross-scenario foundation models), reviews representative methods in each category, and closes with four future directions. The review emphasizes design principles (e.g., data augmentation, representation alignment, prompt/prototype adaptation, feature-space standardization, LLM-based detectors) rather than experimental comparisons.

Significance. The paper addresses a real gap: prior GAD surveys do not focus specifically on generalization. The formal definitions are largely clear, the organization by design principle is intuitive, and the figures convey the paradigm distinctions well. If the corpus is accepted as representative, the taxonomy and problem formalization would help researchers position new work. The paper's strengths include its clear definitions, helpful schematic figures, and an up-to-date reference list; it does not make quantitative predictions or provide code/proofs, so those verification tools are not applicable to a survey. The main caveat is that the survey's central claim of being 'comprehensive' is not verifiable from the manuscript alone, and several taxonomy assignments appear to conflict with the paper's own definitions; these issues are fixable and do not invalidate the underlying organization.

major comments (4)
  1. [II-C and IV-A] Definition 3 defines a GAD foundation model as a scoring function that predicts anomalies for arbitrary samples from an unseen dataset not in the training collection, with both cross-dataset and cross-granularity capability. However, Section IV-A classifies HO-GAT [8], HeagNet [5], BOURNE [14], UniGAD [13], UniFORM [35], SIGNET [20], GRAM [47], and ASD-HC [36] as cross-granularity GAD foundation models without presenting any evidence that these methods are evaluated on datasets outside their training set; as described, they are multi-task detectors for node-, edge-, and graph-level anomalies on a single graph or domain. This mismatch between the formal definition and the taxonomy means the foundation-model category currently includes methods that do not satisfy the paper's own 'unseen dataset' condition. Please either add per-method evidence of cross-dataset evaluation, tighten the definition, or rename the category (e.g., 'multi-granularity GAD').
  2. [III-A.1 and III-B.2] GOODAT [38] is included in the transfer-learning branch, both under generalization-centric training and under test-time fine-tuning. Yet the description states that GOODAT 'repurposes a pre-trained GNN classifier without fine-tuning' and does not mention any source dataset or knowledge transfer from GAD data. This conflicts with Definition 2, which requires 'additional data resources from one or more source datasets.' If GOODAT is meant to be covered, Definition 2 should be broadened to test-time adaptation under distribution shift; otherwise GOODAT belongs to a separate OOD-detection category rather than transfer learning for GAD.
  3. [Introduction and Section II (overall)] The paper's headline claim is that it provides 'a comprehensive and systematic survey of generalization in GAD.' The manuscript does not report a search strategy, database list, query terms, date range, inclusion/exclusion criteria, or screening procedure, so the representativeness of the corpus cannot be audited. This is a load-bearing omission for a survey: without it, the reader cannot distinguish a systematic map from a curated sample, and the taxonomy's completeness is not checkable. Please add a survey-methodology paragraph or appendix that details corpus construction, and consider a table listing each reviewed method with its categorized level and evaluation setting.
  4. [II-C and III] The boundary between the two main paradigms is not enforced by operational criteria. Several methods classified as foundation models (e.g., ARC [21], AnomalyGFM [31], UNPrompt [27]) are also described as using few-shot target data and in-context learning, which functionally resembles transfer learning, while AdaGraph-T3 appears in both transferable-knowledge learning and test-time fine-tuning without an explicit explanation of the overlap. Since the taxonomy is the paper's main contribution, defining a crisp, testable separation (for example, based on whether training data include multiple domains or whether feature spaces are unaligned) and applying it consistently would strengthen the paper.
minor comments (5)
  1. [Abstract] The abstract contains a typo: 'throughtransfer learning' should be 'through transfer learning'.
  2. [Figure 2] The label 'Auxiliary modules and constrains' should read 'Auxiliary modules and constraints'.
  3. [II-C] The paper alternates between 'foundation models' and 'foundational models' (for example, in the taxonomy paragraph); standardize on one term, preferably 'foundation models'.
  4. [III-B] The subsection title 'Test-Time Fine-Tuning' is misleading for GOODAT [38], which is explicitly described as operating without fine-tuning; consider renaming the subsection to 'Test-Time Adaptation' and noting which methods actually fine-tune.
  5. [V] The sentence 'So far, only [1] has explored this issue via heuristic data selection strategies' overstates the contribution of Wild-GAD [1]; the cited paper is a data-selection method rather than a theory of transferability, so the wording should be softened.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the survey's taxonomy and review are organizational, not derivational.

full rationale

This is a literature survey, so its "derivation chain" is taxonomic rather than predictive. The paper formalizes problem settings in Definitions 1-3 and then organizes surveyed methods under the resulting taxonomy. Those definitions do not presuppose the membership of any specific method, and no equation in the paper outputs a quantity that was used as an input, fitted parameter, or renamed prediction. The survey's claim of comprehensiveness rests on an undocumented paper-selection process with no explicit search strategy or inclusion criteria; that is a legitimate auditability and representativeness concern, but it is not circularity under the enumerated patterns. Several surveyed papers are authored by the present authors (e.g., Refs. [20], [21], [28], [31], [41]), but no load-bearing claim is justified exclusively by a self-citation: each method is described by its own published contribution, and the taxonomy itself does not depend on those citations for its validity. There is no imported uniqueness theorem, no ansatz smuggled in via citation, and no known result merely renamed with new coordinates. Accordingly, no specific circular reduction can be exhibited, and the correct finding is no significant circularity.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

As a survey, the paper introduces no fitted parameters or invented entities. Its taxonomic claims rest on the accuracy of its summaries of prior work, on the adequacy of its binary framing, and on the asserted gap in prior surveys.

assumptions (3)
  • domain assumption The survey's summaries of the cited GAD methods are accurate.
    The taxonomy's correctness depends on the fidelity of the paraphrases; we cannot verify every method's description against its original publication.
  • ad hoc to paper Generalization in GAD is adequately divided into transfer learning and foundation models.
    This binary is the paper's proposed framing; it assumes no third paradigm (e.g., continual learning) is needed to organize the surveyed works.
  • domain assumption The prior GAD surveys (e.g., [25], [32]) do not cover generalization in a systematic way, justifying the paper's claimed gap.
    The paper asserts a gap without a detailed comparison to those surveys.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Survey of Generalization of Graph Anomaly Detection: From Transfer Learning to Foundation Models." pith.science (2026). https://pith.science/paper/6S4LSO2D

@misc{pith2026250906609,
  author       = {Pith},
  title        = {Pith review of: A Survey of Generalization of Graph Anomaly Detection: From Transfer Learning to Foundation Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6S4LSO2D}},
  note         = {Machine review of arXiv:2509.06609}
}
read the original abstract

Graph anomaly detection (GAD) has attracted increasing attention in recent years for identifying malicious samples in a wide range of graph-based applications, such as social media and e-commerce. However, most GAD methods assume identical training and testing distributions and are tailored to specific tasks, resulting in limited adaptability to real-world scenarios such as shifting data distributions and scarce training samples in new applications. To address the limitations, recent work has focused on improving the generalization capability of GAD models through transfer learning that leverages knowledge from related domains to enhance detection performance, or developing "one-for-all" GAD foundation models that generalize across multiple applications. Since a systematic understanding of generalization in GAD is still lacking, in this paper, we provide a comprehensive review of generalization in GAD. We first trace the evolution of generalization in GAD and formalize the problem settings, which further leads to our systematic taxonomy. Rooted in this fine-grained taxonomy, an up-to-date and comprehensive review is conducted for the existing generalized GAD methods. Finally, we identify current open challenges and suggest future directions to inspire future research in this emerging field.

Figures

Figures reproduced from arXiv: 2509.06609 by the authors.

Figure 1
Figure 1. The learning paradigms of (a) conventional GAD methods; (b) transfer learning for GAD; and (c) foundation models for GAD. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Sketch maps of transferable knowledge learning. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Sketch maps of target-specific patterns capturing. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Sketch maps of cross-granularity GAD foundation models. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: Sketch maps of cross-scenario GAD foundation models. [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. ProTAGAD: A Foundation Model for TAG Anomaly Detection with Decoupled Topological and Textual Prototypes

    cs.LG 2026-08 conditional novelty 6.0 of 10

    ProTAGAD separates text and graph-structure anomaly evidence into two prototype banks and reports state-of-the-art zero-shot results on eight synthetic text-attributed graph targets.

Reference graph

Works this paper leans on

53 extracted references · 53 canonical work pages · cited by 1 Pith paper

  1. [8]

    Hybrid-order anomaly detection on attributed networks.TKDE, 35(12):12249–12263, 2021

    Ling Huang, Ye Zhu, Yuefang Gao, Tuo Liu, Chao Chang, Caixing Liu, Yong Tang, and Chang-Dong Wang. Hybrid-order anomaly detection on attributed networks.TKDE, 35(12):12249–12263, 2021

    work page 2021

  2. [5]

    Simultaneously detecting node and edge level anomalies on heterogeneous attributed graphs

    Rizal Fathony, Jenn Ng, and Jia Chen. Simultaneously detecting node and edge level anomalies on heterogeneous attributed graphs. InIJCNN, pages 1–10. IEEE, 2024

    work page 2024

  3. [14]

    Bourne: Bootstrapped self-supervised learning framework for unified graph anomaly detection

    Jie Liu, Mengting He, Xuequn Shang, Jieming Shi, Bin Cui, and Hongzhi Yin. Bourne: Bootstrapped self-supervised learning framework for unified graph anomaly detection. InICDE, pages 2820–2833. IEEE, 2024

    work page 2024

  4. [13]

    Vicky Zhao, Yuan Yao, and Jia Li

    Yiqing Lin, Jianheng Tang, Chenyi Zi, H. Vicky Zhao, Yuan Yao, and Jia Li. UniGAD: Unifying multi-level graph anomaly detection. In NeurIPS, 2024

    work page 2024

  5. [35]

    Uniform: Towards unified framework for anomaly detection on graphs

    Chuancheng Song, Xixun Lin, Hanyang Shen, Yanmin Shang, and Yanan Cao. Uniform: Towards unified framework for anomaly detection on graphs. InAAAI, volume 39, pages 12559–12567, 2025

    work page 2025

  6. [20]

    Towards self-interpretable graph-level anomaly detection

    Yixin Liu, Kaize Ding, Qinghua Lu, Fuyi Li, Leo Yu Zhang, and Shirui Pan. Towards self-interpretable graph-level anomaly detection. InNeurIPS, volume 36, pages 8975–8987, 2023

    work page 2023

  7. [47]

    Gram: An interpretable approach for graph anomaly detection using gradient attention maps.Neural Networks, 178, 2024

    Yifei Yang, Peng Wang, Xiaofan He, and Dongmian Zou. Gram: An interpretable approach for graph anomaly detection using gradient attention maps.Neural Networks, 178, 2024

    work page 2024

  8. [36]

    Anomaly subgraph detection through high-order sampling contrastive learning

    Ying Sun, Wenjun Wang, Nannan Wu, and Chunlong Bao. Anomaly subgraph detection through high-order sampling contrastive learning. In IJCAI, pages 2362–2369, 2024

    work page 2024

  9. [38]

    Goodat: towards test-time graph out-of- distribution detection

    Luzhi Wang, Dongxiao He, He Zhang, Yixin Liu, Wenjie Wang, Shirui Pan, Di Jin, and Tat-Seng Chua. Goodat: towards test-time graph out-of- distribution detection. InAAAI, volume 38, pages 15537–15545, 2024

    work page 2024

  10. [21]

    ARC: A generalist graph anomaly detector with in-context learning

    Yixin Liu, Shiyuan Li, Yu Zheng, Qingfeng Chen, Chengqi Zhang, and Shirui Pan. ARC: A generalist graph anomaly detector with in-context learning. InNeurIPS, 2024

    work page 2024

  11. [31]

    Anomalygfm: Graph foundation model for zero/few-shot anomaly detection

    Hezhe Qiao, Chaoxi Niu, Ling Chen, and Guansong Pang. Anomalygfm: Graph foundation model for zero/few-shot anomaly detection. InKDD, 2025

    work page 2025

  12. [27]

    Zero-shot generalist graph anomaly detection with unified neigh- borhood prompts.IJCAI, 2025

    Chaoxi Niu, Hezhe Qiao, Changlu Chen, Ling Chen, and Guansong Pang. Zero-shot generalist graph anomaly detection with unified neigh- borhood prompts.IJCAI, 2025

    work page 2025

Show all 53 references
  1. [1]

    How to use graph data in the wild to help graph anomaly detection? InKDD, page 61–72, 2025

    Yuxuan Cao, Jiarong Xu, Chen Zhao, Jiaan Wang, Carl Yang, Chunping Wang, and Yang Yang. How to use graph data in the wild to help graph anomaly detection? InKDD, page 61–72, 2025

    2025

  2. [2]

    Wirjanto, and Qinmu Peng

    Jiazhen Chen, Sichao Fu, Zhibin Zhang, Zheng Ma, Mingbin Feng, Tony S. Wirjanto, and Qinmu Peng. Towards cross-domain few-shot graph anomaly detection. InICDM, pages 51–60, 2024

    2024

  3. [3]

    Divide and denoise: Empowering simple models for robust semi-supervised node classification against label noise

    Kaize Ding, Xiaoxiao Ma, Yixin Liu, and Shirui Pan. Divide and denoise: Empowering simple models for robust semi-supervised node classification against label noise. InKDD, pages 574–584, 2024

    2024

  4. [4]

    Cross- domain graph anomaly detection.TNNLS, 33(6):2406–2415, 2021

    Kaize Ding, Kai Shu, Xuan Shan, Jundong Li, and Huan Liu. Cross- domain graph anomaly detection.TNNLS, 33(6):2406–2415, 2021

    2021

  5. [6]

    Alleviating structural distribution shift in graph anomaly detection

    Yuan Gao, Xiang Wang, Xiangnan He, Zhenguang Liu, Huamin Feng, and Yongdong Zhang. Alleviating structural distribution shift in graph anomaly detection. InWWW, pages 357–365, 2023

    2023

  6. [7]

    Anomaly detection in the open world: Normality shift detection, explanation, and adaptation

    Dongqi Han, Zhiliang Wang, Wenqi Chen, Kai Wang, Rui Yu, Su Wang, Han Zhang, Zhihua Wang, Minghui Jin, Jiahai Yang, et al. Anomaly detection in the open world: Normality shift detection, explanation, and adaptation. InNDSS, 2023

    2023

  7. [9]

    Assemble your crew: Automatic multi-agent communication topology design via autoregressive graph generation.arXiv, 2025

    Shiyuan Li, Yixin Liu, Qingsong Wen, Chengqi Zhang, and Shirui Pan. Assemble your crew: Automatic multi-agent communication topology design via autoregressive graph generation.arXiv, 2025

    2025

  8. [10]

    Graph anomaly detection with domain-agnostic pre-training and few-shot adaptation

    Xujia Li and Lei Chen. Graph anomaly detection with domain-agnostic pre-training and few-shot adaptation. InICDE, pages 2667–2680. IEEE, 2024

    2024

  9. [11]

    Relevance-aware anomalous users detection in social network via graph neural network

    Yangyang Li, Yipeng Ji, Shaoning Li, Shulong He, Yinhao Cao, Yifeng Liu, Hong Liu, Xiong Li, Jun Shi, and Yangchao Yang. Relevance-aware anomalous users detection in social network via graph neural network. InIJCNN, pages 1–8. IEEE, 2021

    2021

  10. [12]

    Cross- domain graph level anomaly detection.TKDE, 2024

    Zhong Li, Sheng Liang, Jiayang Shi, and Matthijs van Leeuwen. Cross- domain graph level anomaly detection.TKDE, 2024

    2024

  11. [15]

    Pygod: A python library for graph outlier detection.JMLR, 25(141):1–9, 2024

    Kay Liu, Yingtong Dou, Xueying Ding, Xiyang Hu, Ruitong Zhang, Hao Peng, Lichao Sun, and Philip S Yu. Pygod: A python library for graph outlier detection.JMLR, 25(141):1–9, 2024

    2024

  12. [16]

    Bond: Benchmarking unsupervised outlier node detection on static attributed graphs

    Kay Liu, Yingtong Dou, Yue Zhao, Xueying Ding, Xiyang Hu, Ruitong Zhang, Kaize Ding, Canyu Chen, Hao Peng, Kai Shu, et al. Bond: Benchmarking unsupervised outlier node detection on static attributed graphs. InNeurIPS, volume 35, pages 27021–27035, 2022

    2022

  13. [17]

    Anomalyllm: Few-shot anomaly edge detection for dynamic graphs using large language models

    Shuo Liu, Di Yao, Lanting Fang, Zhetao Li, Wenbin Li, Kaiyu Feng, Xiaowen Ji, and Jingping Bi. Anomalyllm: Few-shot anomaly edge detection for dynamic graphs using large language models. InICDM, pages 785–790, 2024

    2024

  14. [18]

    Test-time adaptation on recommender system with data-centric graph transformation.IJCAI, 2025

    Yating Liu, Xin Zheng, Yi Li, and Yanqing Guo. Test-time adaptation on recommender system with data-centric graph transformation.IJCAI, 2025

    2025

  15. [19]

    Self-supervision improves diffusion models for tabular data imputation

    Yixin Liu, Thalaiyasingam Ajanthan, Hisham Husain, and Vu Nguyen. Self-supervision improves diffusion models for tabular data imputation. InCIKM, pages 1513–1522, 2024

    2024

  16. [22]

    Anomaly detection on attributed networks via contrastive self- supervised learning.TNNLS, 33(6):2378–2392, 2021

    Yixin Liu, Zhao Li, Shirui Pan, Chen Gong, Chuan Zhou, and George Karypis. Anomaly detection on attributed networks via contrastive self- supervised learning.TNNLS, 33(6):2378–2392, 2021

    2021

  17. [23]

    Graph- augmented large language model agents: Current progress and future prospects.arXiv, 2025

    Yixin Liu, Guibin Zhang, Kun Wang, Shiyuan Li, and Shirui Pan. Graph- augmented large language model agents: Current progress and future prospects.arXiv, 2025

    2025

  18. [24]

    Grace: Empowering llm-based software vulnerability detection with graph structure and in-context learning.Journal of Systems and Software, 212:112031, 2024

    Guilong Lu, Xiaolin Ju, Xiang Chen, Wenlong Pei, and Zhilong Cai. Grace: Empowering llm-based software vulnerability detection with graph structure and in-context learning.Journal of Systems and Software, 212:112031, 2024

    2024

  19. [25]

    A comprehensive survey on graph anomaly detection with deep learning.TKDE, 35(12):12012–12038, 2021

    Xiaoxiao Ma, Jia Wu, Shan Xue, Jian Yang, Chuan Zhou, Quan Z Sheng, Hui Xiong, and Leman Akoglu. A comprehensive survey on graph anomaly detection with deep learning.TKDE, 35(12):12012–12038, 2021

    2021

  20. [26]

    Blindguard: Safeguarding llm-based multi-agent systems under unknown attacks.arXiv, 2025

    Rui Miao, Yixin Liu, Yili Wang, Xu Shen, Yue Tan, Yiwei Dai, Shirui Pan, and Xin Wang. Blindguard: Safeguarding llm-based multi-agent systems under unknown attacks.arXiv, 2025

    2025

  21. [28]

    A label-free heterophily-guided approach for unsupervised graph fraud detection

    Junjun Pan, Yixin Liu, Xin Zheng, Yizhen Zheng, Alan Wee-Chung Liew, Fuyi Li, and Shirui Pan. A label-free heterophily-guided approach for unsupervised graph fraud detection. InAAAI, volume 39, pages 12443–12451, 2025

    2025

  22. [29]

    Integrating graphs with large language models: Methods and prospects.IEEE Intelligent Systems, 39(1):64–68, 2024

    Shirui Pan, Yizhen Zheng, and Yixin Liu. Integrating graphs with large language models: Methods and prospects.IEEE Intelligent Systems, 39(1):64–68, 2024

    2024

  23. [30]

    Graph anomaly detection via adap- tive test-time representation learning across out-of-distribution domains

    Delaram Pirhayati and Arlei Silva. Graph anomaly detection via adap- tive test-time representation learning across out-of-distribution domains. arXiv, 2025

    2025

  24. [32]

    Deep graph anomaly detection: A survey and new perspectives.TKDE, 37(9):5106–5126, 2025

    Hezhe Qiao, Hanghang Tong, Bo An, Irwin King, Charu Aggarwal, and Guansong Pang. Deep graph anomaly detection: A survey and new perspectives.TKDE, 37(9):5106–5126, 2025

    2025

  25. [33]

    Heterophilic graph invariant learning for out-of-distribution of fraud detection

    Lingfei Ren, Ruimin Hu, Zheng Wang, Yilin Xiao, Dengshi Li, Junhang Wu, Yilong Zang, Jinzhang Hu, and Zijun Huang. Heterophilic graph invariant learning for out-of-distribution of fraud detection. InMM, pages 11032–11040, 2024

    2024

  26. [34]

    Understanding the information propagation effects of communication topologies in llm-based multi-agent systems

    Xu Shen, Yixin Liu, Yiwei Dai, Yili Wang, Rui Miao, Yue Tan, Shirui Pan, and Xin Wang. Understanding the information propagation effects of communication topologies in llm-based multi-agent systems. In EMNLP, 2025

    2025

  27. [37]

    Gadbench: Revisiting and benchmarking supervised graph anomaly detection

    Jianheng Tang, Fengrui Hua, Ziqi Gao, Peilin Zhao, and Jia Li. Gadbench: Revisiting and benchmarking supervised graph anomaly detection. InNeurIPS, volume 36, pages 29628–29653, 2023

    2023

  28. [39]

    Cross-domain graph anomaly detection via anomaly-aware contrastive alignment

    Qizhou Wang, Guansong Pang, Mahsa Salehi, Wray Buntine, and Christopher Leckie. Cross-domain graph anomaly detection via anomaly-aware contrastive alignment. InAAAI, volume 37, pages 4676– 4684, 2023

    2023

  29. [40]

    Open-set graph anomaly detection via normal structure regularisation

    Qizhou Wang, Guansong Pang, Mahsa Salehi, Xiaokun Xia, and Christo- pher Leckie. Open-set graph anomaly detection via normal structure regularisation. InICLR, 2025

    2025

  30. [41]

    Unifying unsupervised graph-level anomaly detection and out-of-distribution detection: A benchmark

    Yili Wang, Yixin Liu, Xu Shen, Chenyu Li, Rui Miao, Kaize Ding, Ying Wang, Shirui Pan, and Xin Wang. Unifying unsupervised graph-level anomaly detection and out-of-distribution detection: A benchmark. In ICLR, 2025

    2025

  31. [42]

    Graph learning under distribution shifts: A comprehensive survey on domain adaptation, out-of-distribution, and continual learning.arXiv, 2024

    Man Wu, Xin Zheng, Qin Zhang, Xiao Shen, Xiong Luo, Xingquan Zhu, and Shirui Pan. Graph learning under distribution shifts: A comprehensive survey on domain adaptation, out-of-distribution, and continual learning.arXiv, 2024

    2024

  32. [43]

    Dynamic weighted learning for unsupervised domain adaptation

    Ni Xiao and Lei Zhang. Dynamic weighted learning for unsupervised domain adaptation. InCVPR, pages 15242–15251, 2021

    2021

  33. [44]

    Relation-aware graph convolutional networks for agent-initiated social e-commerce recommendation

    Fengli Xu, Jianxun Lian, Zhenyu Han, Yong Li, Yujian Xu, and Xing Xie. Relation-aware graph convolutional networks for agent-initiated social e-commerce recommendation. InCIKM, pages 529–538, 2019

    2019

  34. [45]

    Metagad: Meta representation adaptation for few-shot graph anomaly detection

    Xiongxiao Xu, Kaize Ding, Canyu Chen, and Kai Shu. Metagad: Meta representation adaptation for few-shot graph anomaly detection. In DSAA, pages 1–10. IEEE, 2024

    2024

  35. [46]

    Ad- agent: A multi-agent framework for end-to-end anomaly detection

    Tiankai Yang, Junjun Liu, Wingchun Siu, Jiahang Wang, Zhuangzhuang Qian, Chanjuan Song, Cheng Cheng, Xiyang Hu, and Yue Zhao. Ad- agent: A multi-agent framework for end-to-end anomaly detection. arXiv, 2025

    2025

  36. [48]

    Freegad: A training-free yet effective approach for graph anomaly detection

    Yunfeng Zhao, Yixin Liu, Shiyuan Li, Qingfeng Chen, Yu Zheng, and Shirui Pan. Freegad: A training-free yet effective approach for graph anomaly detection. InCIKM, 2025

    2025

  37. [49]

    Test- time graph neural dataset search with generative projection

    Xin Zheng, Wei Huang, Chuan Zhou, Ming Li, and Shirui Pan. Test- time graph neural dataset search with generative projection. InICML, 2025

    2025

  38. [50]

    Online gnn evaluation under test-time graph distribution shifts

    Xin Zheng, Dongjin Song, Qingsong Wen, Bo Du, and Shirui Pan. Online gnn evaluation under test-time graph distribution shifts. InICLR, 2024

    2024

  39. [51]

    Training- free graph anomaly detection: A simple approach via singular value decomposition

    Cheng Zhou, Guangxia Li, Hao Weng, and Yiyu Xiang. Training- free graph anomaly detection: A simple approach via singular value decomposition. InWWW, pages 4196–4205, 2025

    2025

  40. [52]

    Domain generalization: A survey.TPAMI, 45(4):4396–4415, 2022

    Kaiyang Zhou, Ziwei Liu, Yu Qiao, Tao Xiang, and Chen Change Loy. Domain generalization: A survey.TPAMI, 45(4):4396–4415, 2022

    2022

  41. [53]

    Improving generalizability of graph anomaly detection models via data augmentation.TKDE, 35(12):12721–12735, 2023

    Shuang Zhou, Xiao Huang, Ninghao Liu, Huachi Zhou, Fu-Lai Chung, and Long-Kai Huang. Improving generalizability of graph anomaly detection models via data augmentation.TKDE, 35(12):12721–12735, 2023

    2023

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.