Pith. sign in

REVIEW 3 major objections 6 minor 87 references

Top Ten Challenges Towards Agentic Neural Graph Databases

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Graph databases could learn to write and run their own queries; this paper maps the ten problems to solve.

desk verdict A solid, well-organized research agenda for neural graph databases, though the 'agentic' framing is mostly a rebranding and the paper's own neuro-symbolic recommendation undercuts the centrality of neural query execution. read the letter →

arxiv 2501.14224 v1 pith:6AWPNAXH submitted 2025-01-24 cs.AI cs.DBcs.LG

Jiaxin Bai , Zihao Wang , Yukun Zhou , Hang Yin , Weizhi Fei , Qi Hu , Zheye Deng , Jiayang Cheng
show 11 more authors
Tianshi Zheng Hong Ting Tsang Yisen Gao Zhongwei Xie Yufei Li Lixin Fan Binhang Yuan Wei Wang Lei Chen Xiaofang Zhou Yangqiu Song
This is my paper · ORCID
classification cs.AIcs.DBcs.LG
keywords AgenticNeuralGraphDatabasesnetworksqueryansweringabductivereasoningexecutionlargelanguagemodelsdatamanagement
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

Graph databases today store interconnected data but cannot reason over missing or noisy facts; Neural Graph Databases add learned inference but still rely on human-written queries. This paper argues for a third stage, Agentic Neural Graph Databases, in which the system constructs its own queries, executes them by neural computation over graph embeddings, and continuously updates its own knowledge. To make that case, it identifies ten challenges, from semantic units and abductive reasoning through scaling, privacy, and LLM integration. If those challenges are solved, data management shifts from answering predefined questions to autonomously deciding which questions to ask.

What carries the argument

The key object is the Agentic NGDB architecture, a three-part stack: an interface layer that generates queries from a task and context, a learning-and-inference layer that executes those queries as neural computations over node and edge embeddings, and a system layer that keeps compatibility with existing graph databases and performs updates in latent space. The load-bearing mechanism is the division of labour between symbolic query structure and neural execution: the query is a symbolic template, but answers come from learned embeddings, which lets the system answer over incomplete data and, in principle, over query shapes it was not trained to handle. The ten challenges are the points where that mechanism is currently fragile.

What would settle it

Train an Agentic NGDB only on tree-formed queries, then evaluate it on cyclic EFO-1 queries over the same graph; if its answer accuracy collapses while a symbolic graph database answers the same queries exactly, the claim that neural execution can underpin a general query engine is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that the next step for graph data management is an Agentic NGDB, an architecture whose defining properties are autonomous query construction, neural query execution, and continuous learning. It builds on the earlier NGDB idea of coupling graph neural networks with graph storage, but moves the human out of the query-writing loop. The contribution is the architecture plus an inventory of ten obstacles, grouped by interface, learning, and system; each obstacle is tied to existing results. The paper's stated position is that a database built this way would be an intelligent, self-improving system for modern data-driven applications.

Load-bearing premise

The load-bearing premise is that neural modules can be made reliable and scalable enough to serve as a database execution engine; the paper admits in Section 4 that even for a fixed query family these modules still lose performance when generalizing, and if that gap cannot be closed the Agentic NGDB architecture loses its foundation.

Editorial extensions

If this is right

  • A user could pose a high-level task rather than a precise query, and the database would derive the appropriate query for the context.
  • Answers would remain available even when the stored graph is incomplete or noisy, because the neural execution engine predicts rather than only matches.
  • The system could modify its own stored knowledge through CREATE, UPDATE, and DELETE operations, making continuous self-improvement a database feature.
  • Combined with LLMs, the architecture would support retrieval-augmented generation over structured graph reasoning rather than text similarity alone.
  • The ordering of the ten challenges suggests that interface-level semantics and abductive reasoning must mature before system-level scaling and distribution can succeed.

Reading between the lines

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

  • Beyond the paper: the proposal would be testable through a benchmark that scores whether a system can generate the right query from a natural-language goal on unseen query families, something the paper does not define.
  • A further consequence the authors leave implicit is a cost comparison: the viability of Agentic NGDBs will depend not only on answer accuracy but on where neural execution is cheaper than symbolic execution.
  • The ten challenges could be reordered by dependency; the paper does not state which challenges must be solved before others become tractable.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 6 minor

Summary. The paper proposes a new class of systems, Agentic Neural Graph Databases (Agentic NGDBs), defined as NGDBs extended with three core functionalities: autonomous query construction, neural query execution, and continuous learning. It identifies ten challenges across interface, learning/inference, and system perspectives, and discusses each challenge with reference to prior work, a substantial portion of which is authored by the same research group. The only technical formula is a joint training objective, L_total = L_LLM + lambda L_NGDB, in Section 10. The paper is a vision/position paper without experiments or formal analysis, and it concludes with expected applications in autonomous data management, personalized recommendation, and complex event processing.

Significance. If the research agenda is adopted, the paper could provide a useful organizing framework for a community that spans knowledge-graph reasoning and database systems. Its strength is that it maps a broad set of open problems---semantic units, abductive reasoning, query generalization, privacy, scaling, distributed systems, compatibility, grounding, and LLM integration---onto a single architecture, and it connects recent complex query answering results to systems concerns such as hybrid storage and elastic scalability. It also names concrete gaps, including the lack of privacy evaluation benchmarks and the performance loss of neural modules even for fixed query families. On the other hand, the paper contains no experiments, no formal definitions, no reproducible code, and no machine-checked proofs; its only equation is a generic linear loss combination. The contribution is therefore entirely in the coherence and precision of its agenda, and in that respect the internal tension around the role of neural query execution is material.

major comments (3)
  1. [Section 4.2 and Section 1] The central functionality 'neural query execution' is internally inconsistent with the paper's recommended generalization strategy. Section 1 defines Agentic NGDBs as extending NGDBs with 'neural query execution' as one of three core functionalities, and Section 4.1 notes that neural modules 'still suffer a loss in performance for generalization even when the query family is fixed' [16]. Yet Section 4.2 recommends deriving an instance of a classic graph database using a link predictor and then applying standard graph-database query processing, claiming 'the neuro-symbolic approach achieves the same level of generalizability in queries as the classic database research.' If the reliable route is neuro-symbolic, the neural component is a data-completion preprocessor and query execution is symbolic; if the intended route is end-to-end neural execution (query embeddings, sequence models), the paper concedes a persistent generalization gap and provides no evidence that it can be closed. The authors must define 'neural query execution' precisely, state which execution mode is required for each of the three core functionalities, and reconcile these statements.
  2. [Abstract and Section 1] The paper's title and abstract claim these are the 'top ten' challenges and the 'most critical' challenges, but no methodology, criteria, or comparative evidence is provided to justify this ranking. The list is selected 'based on recent progress in the research community' (Section 1), yet many challenges are motivated by the authors' own prior work (e.g., [11,14,15,16,25,26,30,42,48,49]), which makes the selection difficult to verify independently. Without a stated selection procedure, the central organizational claim of the paper is unfalsifiable. I recommend either adding a short methodology subsection that explains how the list was compiled or softening the claim to 'a set of challenges' and removing 'Top Ten' from the title.
  3. [Section 10, Eq. (1)] The joint learning framework for LLM-NGDB integration is presented as the combined objective L_total = L_LLM + lambda L_NGDB, but this equation is a generic linear combination and does not specify how the three Agentic NGDB functionalities (autonomous query construction, neural query execution, continuous learning) enter the loss, how parameters are shared, or how the feedback loop is trained. No experiments or concrete instantiations are provided, so it is unclear whether Eq. (1) is a proposal, a definition, or a result. Because Challenge 9 is one of the ten central challenges, this under-specification leaves a load-bearing part of the agenda unevaluable.
minor comments (6)
  1. [Section 7.1] The list item 'Scalability..' contains a double period and should be 'Scalability.'
  2. [Section 5.2] The heading 'Private Distribute Training' is a typo; it should be 'Private Distributed Training.'
  3. [Section 3] The bullet items 'More Generalized Observation', 'More Complex Structured Hypotheses', 'Graph-Based Hypothesis Generation Models', and 'NGDB as a Reward Model' are sentence fragments; each should be a complete sentence or a consistent noun phrase with punctuation.
  4. [Section 4.1] The phrase 'The set operations composition allows the models to generalize the entire tree-form query family' is grammatically awkward, and 'tree-form' appears alongside 'tree-formed' in the same paragraph; the terminology should be made consistent.
  5. [Section 2] The examples 'It is raining outside' and 'This movie is great' are informal and would benefit from quotation marks or italics to mark them as example beliefs.
  6. [Figure 1 and Section 11] Challenge 10 is framed as a research challenge but the section describes application domains rather than an open technical problem; this is inconsistent with the other nine items and should either be reworked as a challenge or reframed as 'applications and their requirements.'

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the paper proposes a definition and a research agenda, with no derived predictions that reduce to their inputs.

full rationale

The paper's central contribution is the definition of Agentic NGDBs and a curated list of ten research challenges. The three core functionalities (autonomous query construction, neural query execution, continuous learning) are introduced as definitions, not derived from any prior result, so they cannot be circular in the sense of one result being equivalent to its input. The ten challenges are selected based on existing literature and the authors' own prior work, but this is a research-prioritization choice, not an empirical prediction. The self-citations (e.g., [11], [15], [16], [25], [26], [30], [42], [48], [49]) are used to ground claims about prior methods, benchmarks, and known limitations; they are concrete, externally checkable works rather than unverified uniqueness theorems or fitted parameters. The only equation in the paper, Ltotal = LLLM + λLNGDB (Section 10, Challenge 9), is a proposed joint-training objective, not a derivation of Agentic NGDB capabilities from assumptions that already contain them. The paper also contains an explicit feasibility limitation: Section 4 states that neural modules 'still suffer a loss in performance for generalization even when the query family is fixed' (citing [16]), and Section 4.2 recommends a neuro-symbolic reduction that 'achieves the same level of generalizability in queries as the classic database research.' That reduction is stated transparently as 'derive an instance of a classic graph database using the link predictor, and all previous research in graph databases applies directly'; it is an explicit design rationale, not a hidden circularity. The tension between 'neural query execution' as a core functionality and the admitted reliability gap of end-to-end neural query execution is a coherence and feasibility concern for the proposed vision, but it is not a circularity: no fitted value is renamed as a prediction, no definition presupposes the target conclusion, and no argument reduces by construction to a self-citation. Therefore the appropriate circularity score is 0.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The paper's central proposal rests on definitional and domain assumptions rather than derived results. The main free parameters are absent because no model is fit. The axioms capture the framing choices: the claimed lack of autonomy in NGDBs, the necessity of the three core functionalities, and the selection of the ten challenges. The invented entity is the Agentic NGDB concept itself, which is proposed as a target architecture without implementation or independent evidence.

assumptions (4)
  • domain assumption NGDBs rely on predefined queries and lack autonomy and adaptability.
    Stated in the abstract and Section 1; this motivates the entire Agentic NGDB concept but is not empirically established in the paper.
  • ad hoc to paper The three core functionalities (autonomous query construction, neural query execution, continuous learning) are the necessary extensions of NGDBs.
    Proposed by the authors without derivation or comparison to alternative decompositions.
  • ad hoc to paper The ten listed challenges are the most critical ones for realizing Agentic NGDBs.
    The selection is presented in Section 1 and reflected in Figure 1, but no systematic survey or ranking methodology is given.
  • domain assumption Neural query execution can eventually match symbolic execution for practical query families.
    Underlies the whole neural execution premise; Section 4 admits neural modules still lose performance even on fixed query families (citing [16]).
invented entities (1)
  • Agentic NGDB (Agentic Neural Graph Database)
    purpose: Serves as the proposed target architecture: a graph database with autonomous query construction, neural query execution, and continuous learning.
    The concept is defined in this paper and not instantiated or benchmarked anywhere; the paper explicitly says future work will develop the co-training framework and integration modules.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Top Ten Challenges Towards Agentic Neural Graph Databases." pith.science (2026). https://pith.science/paper/6AWPNAXH

@misc{pith2026250114224,
  author       = {Pith},
  title        = {Pith review of: Top Ten Challenges Towards Agentic Neural Graph Databases},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6AWPNAXH}},
  note         = {Machine review of arXiv:2501.14224}
}
read the original abstract

Graph databases (GDBs) like Neo4j and TigerGraph excel at handling interconnected data but lack advanced inference capabilities. Neural Graph Databases (NGDBs) address this by integrating Graph Neural Networks (GNNs) for predictive analysis and reasoning over incomplete or noisy data. However, NGDBs rely on predefined queries and lack autonomy and adaptability. This paper introduces Agentic Neural Graph Databases (Agentic NGDBs), which extend NGDBs with three core functionalities: autonomous query construction, neural query execution, and continuous learning. We identify ten key challenges in realizing Agentic NGDBs: semantic unit representation, abductive reasoning, scalable query execution, and integration with foundation models like large language models (LLMs). By addressing these challenges, Agentic NGDBs can enable intelligent, self-improving systems for modern data-driven applications, paving the way for adaptable and autonomous data management solutions.

Figures

Figures reproduced from arXiv: 2501.14224 by the authors.

Figure 1
Figure 1. The top ten challenges in achieving Agentic NGDB. Its three perspectives include interface, learning, [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

87 extracted references · 63 canonical work pages

  1. [16]

    Zihao Wang, Hang Yin, and Yangqiu Song. Benchmarking the Combinatorial Generalizability of Complex Query Answering on Knowledge Graphs.Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks, 1, December 2021

    work page 2021

  2. [1]

    Graph database applications and concepts with neo4j

    Justin J Miller. Graph database applications and concepts with neo4j. In Proceedings of the southern association for information systems conference, Atlanta, GA, USA, volume 2324, pages 141–147, 2013

    2013

  3. [2]

    Alin Deutsch, Yu Xu, Mingxi Wu, and Victor E. Lee. Tigergraph: A native mpp graph database. ArXiv, abs/1901.08248, 2019

    arXiv 1901

  4. [3]

    Neural graph databases

    Maciej Besta, Patrick Iff, Florian Scheidl, Kazuki Osawa, Nikoli Dryden, Michal Podstawski, Tiancheng Chen, and Torsten Hoefler. Neural graph databases. In Bastian Rieck and Razvan Pascanu, editors,Learn- ing on Graphs Conference, LoG 2022, 9-12 December 2022, Virtual Event, volume 198 of Proceedings of Machine Learning Research, page 31. PMLR, 2022

    2022

  5. [4]

    Neural graph reason- ing: Complex logical query answering meets graph databases

    Hongyu Ren, Mikhail Galkin, Michael Cochez, Zhaocheng Zhu, and Jure Leskovec. Neural graph reason- ing: Complex logical query answering meets graph databases. arXiv preprint arXiv:2303.14617, 2023

    arXiv 2023

  6. [5]

    Artificial intelligence: a modern approach

    Stuart J Russell and Peter Norvig. Artificial intelligence: a modern approach. Pearson, 2016

    2016

  7. [6]

    Kblrn: End-to-end learning of knowledge base representations with latent, relational, and numerical features

    Alberto Garc ´ıa-Dur´an and Mathias Niepert. Kblrn: End-to-end learning of knowledge base representations with latent, relational, and numerical features. In Amir Globerson and Ricardo Silva, editors, Proceedings of the Thirty-Fourth Conference on Uncertainty in Artificial Intelligence, UAI 2018, Monterey, California, USA, August 6-10, 2018, pages 372–381...

    2018

  8. [7]

    Knowledge representation learning with entities, attributes and relations

    Yankai Lin, Zhiyuan Liu, and Maosong Sun. Knowledge representation learning with entities, attributes and relations. In Subbarao Kambhampati, editor, Proceedings of the Twenty-Fifth International Joint Con- ference on Artificial Intelligence, IJCAI 2016, New York, NY, USA, 9-15 July 2016 , pages 2866–2872. IJCAI/AAAI Press, 2016

    2016

Show all 87 references
  1. [8]

    Litcqd: Multi-hop reasoning in incomplete knowledge graphs with numeric literals

    Caglar Demir, Michel Wiebesiek, Renzhong Lu, Axel-Cyrille Ngonga Ngomo, and Stefan Heindorf. Litcqd: Multi-hop reasoning in incomplete knowledge graphs with numeric literals. In Danai Koutra, Claudia Plant, Manuel Gomez Rodriguez, Elena Baralis, and Francesco Bonchi, editors, ...

    2023

  2. [9]

    Smith, and Yejin Choi

    Maarten Sap, Ronan Le Bras, Emily Allaway, Chandra Bhagavatula, Nicholas Lourie, Hannah Rashkin, Brendan Roof, Noah A. Smith, and Yejin Choi. ATOMIC: an atlas of machine commonsense for if- then reasoning. In The Thirty-Third AAAI Conference on Artificial Intelligence, AAAI 20...

    2019

  3. [10]

    ASER: A large-scale eventuality knowledge graph

    Hongming Zhang, Xin Liu, Haojie Pan, Yangqiu Song, and Cane Wing-Ki Leung. ASER: A large-scale eventuality knowledge graph. In Yennun Huang, Irwin King, Tie-Yan Liu, and Maarten van Steen, editors, WWW ’20: The Web Conference 2020, Taipei, Taiwan, April 20-24, 2020, pages 201–...

    2020

  4. [11]

    Complex query answering on eventuality knowledge graph with implicit logical constraints

    Jiaxin Bai, Xin Liu, Weiqi Wang, Chen Luo, and Yangqiu Song. Complex query answering on eventuality knowledge graph with implicit logical constraints. In Alice Oh, Tristan Naumann, Amir Globerson, Kate Saenko, Moritz Hardt, and Sergey Levine, editors, Advances in Neural Inform...

    2023

  5. [12]

    FolkScope: Intention knowledge graph construction for E-commerce commonsense discovery

    Changlong Yu, Weiqi Wang, Xin Liu, Jiaxin Bai, Yangqiu Song, Zheng Li, Yifan Gao, Tianyu Cao, and Bing Yin. FolkScope: Intention knowledge graph construction for E-commerce commonsense discovery. In Findings of the Association for Computational Linguistics: ACL 2023 , pages 11...

    2023

  6. [13]

    Cosmo: A large-scale e-commerce common sense knowledge generation and serving system at amazon

    Changlong Yu, Xin Liu, Jefferson Maia, Tianyu Cao, Laurence (Yang) Li, Yifan Gao, Yangqiu Song, Rahul Goutam, Haiyang Zhang, Bing Yin, and Zheng Li. Cosmo: A large-scale e-commerce common sense knowledge generation and serving system at amazon. 2024

    2024

  7. [14]

    Intention knowledge graph construction for user intention relation modeling

    Jiaxin Bai, Zhaobo Wang, Junfei Cheng, Dan Yu, Zerui Huang, Weiqi Wang, Xin Liu, Chen Luo, Qi He, Yanming Zhu, et al. Intention knowledge graph construction for user intention relation modeling. arXiv preprint arXiv:2412.11500, 2024

    2024

  8. [15]

    Advancing abductive reasoning in knowledge graphs through complex logical hypothesis generation

    Jiaxin Bai, Yicheng Wang, Tianshi Zheng, Yue Guo, Xin Liu, and Yangqiu Song. Advancing abductive reasoning in knowledge graphs through complex logical hypothesis generation. In In Proceedings of the 62st Annual Meeting of the Association for Computational Linguistics. Associat...

    2024

  9. [17]

    Logical queries on knowledge graphs: Emerging interface of incomplete relational data

    Zihao Wang, Hang Yin, and Yangqiu Song. Logical queries on knowledge graphs: Emerging interface of incomplete relational data. IEEE Data Eng. Bull., 2022

    2022

  10. [18]

    Neural-symbolic reasoning over knowledge graphs: A survey from a query perspective

    Lihui Liu, Zihao Wang, and Hanghang Tong. Neural-symbolic reasoning over knowledge graphs: A survey from a query perspective. arXiv preprint arXiv:2412.10390, 2024

    2024 arXiv

  11. [19]

    Embedding logical queries on knowledge graphs

    Will Hamilton, Payal Bajaj, Marinka Zitnik, Dan Jurafsky, and Jure Leskovec. Embedding logical queries on knowledge graphs. Advances in neural information processing systems, 31, 2018

    2018

  12. [20]

    Query2box: Reasoning Over Knowledge Graphs In Vector Space Using Box Embeddings

    H Ren, W Hu, and J Leskovec. Query2box: Reasoning Over Knowledge Graphs In Vector Space Using Box Embeddings. In International Conference on Learning Representations (ICLR), 2020

    2020

  13. [21]

    Neural-Answering Logical Queries on Knowledge Graphs

    Lihui Liu, Boxin Du, Heng Ji, ChengXiang Zhai, and Hanghang Tong. Neural-Answering Logical Queries on Knowledge Graphs. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, pages 1087–1097, 2021

    2021

  14. [22]

    Beta embeddings for multi-hop logical reasoning in knowledge graphs

    Hongyu Ren and Jure Leskovec. Beta embeddings for multi-hop logical reasoning in knowledge graphs. Advances in Neural Information Processing Systems, 33:19716–19726, 2020

    2020

  15. [23]

    Wasserstein-fisher-rao embedding: Logical query embeddings with local comparison and global transport

    Zihao Wang, Weizhi Fei, Hang Yin, Yangqiu Song, Ginny Wong, and Simon See. Wasserstein-fisher-rao embedding: Logical query embeddings with local comparison and global transport. In Findings of the Association for Computational Linguistics: ACL 2023, pages 13679–13696, 2023

    2023

  16. [24]

    Meta operator for complex query answering on knowledge graphs

    Hang Yin, Zihao Wang, and Yangqiu Song. Meta operator for complex query answering on knowledge graphs. arXiv preprint arXiv:2403.10110, 2024

    2024 arXiv

  17. [25]

    Rethinking existential first order queries and their inference on knowledge graphs

    Hang Yin, Zihao Wang, and Yangqiu Song. Rethinking existential first order queries and their inference on knowledge graphs. In The Twelfth International Conference on Learning Representations, 2024

    2024

  18. [26]

    Logical message passing networks with one-hop inference on atomic formulas

    Zihao Wang, Yangqiu Song, Ginny Y Wong, and Simon See. Logical message passing networks with one-hop inference on atomic formulas. arXiv preprint arXiv:2301.08859, 2023. 14

    2023 arXiv

  19. [27]

    NQE: N-ary Query Embedding for Complex Query Answering over Hyper-relational Knowledge Graphs

    Haoran Luo, Yuhao Yang, Gengxian Zhou, Yikai Guo, Tianyu Yao, Zichen Tang, Xueyuan Lin, Kaiyang Wan, and others. NQE: N-ary Query Embedding for Complex Query Answering over Hyper-relational Knowledge Graphs. arXiv preprint arXiv:2211.13469, 2022

    2022 arXiv

  20. [28]

    Knowledge graph embedding: A survey of ap- proaches and applications

    Quan Wang, Zhendong Mao, Bin Wang, and Li Guo. Knowledge graph embedding: A survey of ap- proaches and applications. IEEE transactions on knowledge and data engineering , 29(12):2724–2743, 2017

    2017

  21. [29]

    Answering Complex Logical Queries on Knowledge Graphs via Query Computation Tree Optimization

    Yushi Bai, Xin Lv, Juanzi Li, and Lei Hou. Answering Complex Logical Queries on Knowledge Graphs via Query Computation Tree Optimization. In Proceedings of the 40th International Conference on Machine Learning, pages 1472–1491. PMLR, July 2023. ISSN: 2640-3498

    2023

  22. [30]

    Sequential query encoding for complex query answering on knowledge graphs

    Jiaxin Bai, Tianshi Zheng, and Yangqiu Song. Sequential query encoding for complex query answering on knowledge graphs. Transactions on Machine Learning Research, 2023

    2023

  23. [31]

    Database privacy: balancing confidentiality, integrity and availability

    Martin S Olivier. Database privacy: balancing confidentiality, integrity and availability. ACM SIGKDD Explorations Newsletter, 4(2):20–27, 2002

    2002

  24. [32]

    Access control for databases: Concepts and systems

    Elisa Bertino, Gabriel Ghinita, Ashish Kamra, et al. Access control for databases: Concepts and systems. Foundations and Trends® in Databases, 3(1–2):1–148, 2011

    2011

  25. [33]

    Detecting insider threats in a real corporate database of computer usage activity

    Ted E Senator, Henry G Goldberg, Alex Memory, William T Young, Brad Rees, Robert Pierce, Daniel Huang, Matthew Reardon, David A Bader, Edmond Chow, et al. Detecting insider threats in a real corporate database of computer usage activity. In Proceedings of the 19th ACM SIGKDD i...

    2013

  26. [34]

    Inference attacks on property-preserving en- crypted databases

    Muhammad Naveed, Seny Kamara, and Charles V Wright. Inference attacks on property-preserving en- crypted databases. In Proceedings of the 22nd ACM SIGSAC Conference on Computer and Communica- tions Security, pages 644–655, 2015

    2015

  27. [35]

    A comparative study of data anonymization techniques

    Suntherasvaran Murthy, Asmidar Abu Bakar, Fiza Abdul Rahim, and Ramona Ramli. A comparative study of data anonymization techniques. In 2019 IEEE 5th Intl Conference on Big Data Security on Cloud (BigDataSecurity), IEEE Intl Conference on High Performance and Smart Computing,(H...

    2019

  28. [36]

    k-anonymity: A model for protecting privacy

    Latanya Sweeney. k-anonymity: A model for protecting privacy. International journal of uncertainty, fuzziness and knowledge-based systems, 10(05):557–570, 2002

    2002

  29. [37]

    l-diversity: Privacy beyond k-anonymity

    Ashwin Machanavajjhala, Daniel Kifer, Johannes Gehrke, and Muthuramakrishnan Venkitasubramaniam. l-diversity: Privacy beyond k-anonymity. Acm transactions on knowledge discovery from data (tkdd) , 1(1):3–es, 2007

    2007

  30. [38]

    Database security and encryption: A survey study

    Iqra Basharat, Farooque Azam, and Abdul Wahab Muzaffar. Database security and encryption: A survey study. International Journal of Computer Applications, 47(12), 2012

    2012

  31. [39]

    Differential privacy

    Cynthia Dwork. Differential privacy. In International colloquium on automata, languages, and program- ming, pages 1–12. Springer, 2006

    2006

  32. [40]

    Adversarial attacks on link prediction algorithms based on graph neural networks

    Wanyu Lin, Shengxiang Ji, and Baochun Li. Adversarial attacks on link prediction algorithms based on graph neural networks. In Proceedings of the 15th ACM Asia Conference on Computer and Communica- tions Security, pages 370–380, 2020. 15

    2020

  33. [41]

    Structural attack against graph based android malware detection

    Kaifa Zhao, Hao Zhou, Yulin Zhu, Xian Zhan, Kai Zhou, Jianfeng Li, Le Yu, Wei Yuan, and Xiapu Luo. Structural attack against graph based android malware detection. InProceedings of the 2021 ACM SIGSAC conference on computer and communications security, pages 3218–3235, 2021

    2021

  34. [42]

    Privacy-preserved neural graph databases

    Qi Hu, Haoran Li, Jiaxin Bai, Zihao Wang, and Yangqiu Song. Privacy-preserved neural graph databases. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages 1108–1118, 2024

    2024

  35. [43]

    Privacy leakage on dnns: A survey of model inversion attacks and defenses

    Hao Fang, Yixiang Qiu, Hongyao Yu, Wenbo Yu, Jiawei Kong, Baoli Chong, Bin Chen, Xuan Wang, Shu- Tao Xia, and Ke Xu. Privacy leakage on dnns: A survey of model inversion attacks and defenses. arXiv preprint arXiv:2402.04013, 2024

    2024 arXiv

  36. [44]

    Membership inference attacks on machine learning: A survey

    Hongsheng Hu, Zoran Salcic, Lichao Sun, Gillian Dobbie, Philip S Yu, and Xuyun Zhang. Membership inference attacks on machine learning: A survey. ACM Computing Surveys (CSUR), 54(11s):1–37, 2022

    2022

  37. [45]

    Information leakage in embedding models

    Congzheng Song and Ananth Raghunathan. Information leakage in embedding models. In Proceedings of the 2020 ACM SIGSAC conference on computer and communications security, pages 377–390, 2020

    2020

  38. [46]

    Effects of differential privacy and data skewness on membership inference vulnerability

    Stacey Truex, Ling Liu, Mehmet Emre Gursoy, Wenqi Wei, and Lei Yu. Effects of differential privacy and data skewness on membership inference vulnerability. In2019 First IEEE international conference on trust, privacy and security in intelligent systems and applications (TPS-IS...

    2019

  39. [47]

    Regression model fitting under differential privacy and model inversion attack

    Yue Wang, Cheng Si, and Xintao Wu. Regression model fitting under differential privacy and model inversion attack. In Twenty-fourth international joint conference on artificial intelligence, 2015

    2015

  40. [48]

    Independent distribution regularization for private graph embedding

    Qi Hu and Yangqiu Song. Independent distribution regularization for private graph embedding. In Pro- ceedings of the 32nd ACM International Conference on Information and Knowledge Management , pages 823–832, 2023

    2023

  41. [49]

    Fedcqa: Answering complex queries on multi-source knowledge graphs via federated learning

    Qi Hu, Weifeng Jiang, Haoran Li, Zihao Wang, Jiaxin Bai, Qianren Mao, Yangqiu Song, Lixin Fan, and Jianxin Li. Fedcqa: Answering complex queries on multi-source knowledge graphs via federated learning. CoRR, 2024

    2024

  42. [50]

    Secure multi-party computation

    Oded Goldreich. Secure multi-party computation. Manuscript. Preliminary version, 78(110):1–108, 1998

    1998

  43. [51]

    A survey on homomorphic encryption schemes: Theory and implementation

    Abbas Acar, Hidayet Aksu, A Selcuk Uluagac, and Mauro Conti. A survey on homomorphic encryption schemes: Theory and implementation. ACM Computing Surveys (Csur), 51(4):1–35, 2018

    2018

  44. [52]

    Deep learning scaling is predictable, empirically

    Joel Hestness, Sharan Narang, Newsha Ardalani, Gregory Diamos, Heewoo Jun, Hassan Kianinejad, Md Mostofa Ali Patwary, Yang Yang, and Yanqi Zhou. Deep learning scaling is predictable, empirically. arXiv preprint arXiv:1712.00409, 2017

    2017 arXiv

  45. [53]

    Scaling laws for neural language models

    Jared Kaplan, Sam McCandlish, Tom Henighan, Tom B Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amodei. Scaling laws for neural language models. arXiv preprint arXiv:2001.08361, 2020

    2001 arXiv

  46. [54]

    $ \text{EFO} {k}$-CQA: Towards Knowledge Graph Complex Query Answering beyond Set Operation, July 2023

    Hang Yin, Zihao Wang, Weizhi Fei, and Yangqiu Song. $ \text{EFO} {k}$-CQA: Towards Knowledge Graph Complex Query Answering beyond Set Operation, July 2023. arXiv:2307.13701 [cs]

    2023 arXiv

  47. [55]

    Smore: Knowledge graph completion and multi-hop reasoning in massive knowledge graphs

    Hongyu Ren, Hanjun Dai, Bo Dai, Xinyun Chen, Denny Zhou, Jure Leskovec, and Dale Schuurmans. Smore: Knowledge graph completion and multi-hop reasoning in massive knowledge graphs. In Proceed- ings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages ...

  48. [56]

    Foundations of databases, volume 8

    Serge Abiteboul, Richard Hull, and Victor Vianu. Foundations of databases, volume 8. Addison-Wesley Reading, 1995

    1995

  49. [57]

    Freebase: a collaboratively created graph database for structuring human knowledge

    Kurt Bollacker, Colin Evans, Praveen Paritosh, Tim Sturge, and Jamie Taylor. Freebase: a collaboratively created graph database for structuring human knowledge. In Proceedings of the 2008 ACM SIGMOD international conference on Management of data, pages 1247–1250, 2008

    2008

  50. [58]

    Knowledge graphs querying

    Arijit Khan. Knowledge graphs querying. ACM SIGMOD Record, 52(2):18–29, 2023

    2023

  51. [59]

    The world of graph databases from an industry perspective

    Yuanyuan Tian. The world of graph databases from an industry perspective. ACM SIGMOD Record , 51(4):60–67, 2023

    2023

  52. [60]

    Manu: a cloud native vector database management system

    Rentong Guo, Xiaofan Luan, Long Xiang, Xiao Yan, Xiaomeng Yi, Jigao Luo, Qianya Cheng, Weizhi Xu, Jiarui Luo, Frank Liu, et al. Manu: a cloud native vector database management system. arXiv preprint arXiv:2206.13843, 2022

    2022 arXiv

  53. [61]

    Survey of vector database management systems

    James Jie Pan, Jianguo Wang, and Guoliang Li. Survey of vector database management systems. The VLDB Journal, 33(5):1591–1615, 2024

    2024

  54. [62]

    Graph database partitioning: A study

    Ali Ben Ammar. Graph database partitioning: A study. In 2016 7th International Conference on Informa- tion, Intelligence, Systems & Applications (IISA), pages 1–9. IEEE, 2016

    2016

  55. [63]

    Geabase: A high-performance distributed graph database for industry-scale applications

    Zhisong Fu, Zhengwei Wu, Houyi Li, Yize Li, Min Wu, Xiaojie Chen, Xiaomeng Ye, Benquan Yu, and Xi Hu. Geabase: A high-performance distributed graph database for industry-scale applications. Interna- tional Journal of High Performance Computing and Networking, 15(1-2):12–21, 2019

    2019

  56. [64]

    Bytegraph: a high-performance distributed graph database in bytedance

    Changji Li, Hongzhi Chen, Shuai Zhang, Yingqian Hu, Chao Chen, Zhenjie Zhang, Meng Li, Xiangchen Li, Dongqing Han, Xiaohui Chen, et al. Bytegraph: a high-performance distributed graph database in bytedance. Proceedings of the VLDB Endowment, 15(12):3306–3318, 2022

    2022

  57. [65]

    Design of highly scalable graph database systems without exponential perfor- mance degradation

    Ricky Sun and Jamie Chen. Design of highly scalable graph database systems without exponential perfor- mance degradation. In Proceedings of the International Workshop on Big Data in Emergent Distributed Environments, pages 1–6, 2023

    2023

  58. [66]

    Inc. Neo4j. Neo4j. https://neo4j.com, 2023. Accessed: 2024-10-28

    2023

  59. [67]

    TigerGraph

    Inc. TigerGraph. Tigergraph. https://www.tigergraph.com, 2023. Accessed: 2024-10-28

    2023

  60. [68]

    Milvus: A purpose-built vector data management system

    Jianguo Wang, Xiaomeng Yi, Rentong Guo, Hai Jin, Peng Xu, Shengjun Li, Xiangyu Wang, Xiangzhou Guo, Chengming Li, Xiaohai Xu, et al. Milvus: A purpose-built vector data management system. In Proceedings of the 2021 International Conference on Management of Data, pages 2614–2627, 2021

    2021

  61. [69]

    Qdrant Team. Qdrant. https://qdrant.tech, 2023. Accessed: 2024-10-28

    2023

  62. [70]

    Weaviate

    SeMI Technologies. Weaviate. https://weaviate.io, 2023. Accessed: 2024-10-28

    2023

  63. [71]

    A survey on distributed graph pattern matching in massive graphs

    Sarra Bouhenni, Said Yahiaoui, Nadia Nouali-Taboudjemat, and Hamamache Kheddouci. A survey on distributed graph pattern matching in massive graphs. ACM Computing Surveys (CSUR) , 54(2):1–35, 2021

    2021

  64. [72]

    A survey of distributed graph algorithms on massive graphs

    Lingkai Meng, Yu Shao, Long Yuan, Longbin Lai, Peng Cheng, Xue Li, Wenyuan Yu, Wenjie Zhang, Xuemin Lin, and Jingren Zhou. A survey of distributed graph algorithms on massive graphs. ACM Com- puting Surveys, 57(2):1–39, 2024. 17

    2024

  65. [73]

    Atom: An efficient query serving system for embedding-based knowledge graph reasoning with operator-level batching

    Qihui Zhou, Peiqi Yin, Xiao Yan, Changji Li, Guanxian Jiang, and James Cheng. Atom: An efficient query serving system for embedding-based knowledge graph reasoning with operator-level batching. Proceed- ings of the ACM on Management of Data, 2(4):1–29, 2024

    2024

  66. [74]

    nsdb: Architecting the next generation database by integrating neural and symbolic systems.Proceedings of the VLDB Endowment, 17(11):3283– 3289, 2024

    Ye Yuan, Bo Tang, Tianfei Zhou, Zhiwei Zhang, and Jianbin Qin. nsdb: Architecting the next generation database by integrating neural and symbolic systems.Proceedings of the VLDB Endowment, 17(11):3283– 3289, 2024

    2024

  67. [75]

    Neurdb: On the design and implementation of an ai-powered autonomous database

    Zhanhao Zhao, Shaofeng Cai, Haotian Gao, Hexiang Pan, Siqi Xiang, Naili Xing, Gang Chen, Beng Chin Ooi, Yanyan Shen, Yuncheng Wu, et al. Neurdb: On the design and implementation of an ai-powered autonomous database. arXiv preprint arXiv:2408.03013, 2024

    2024 arXiv

  68. [76]

    Optimizing llm queries in relational workloads

    Shu Liu, Asim Biswal, Audrey Cheng, Xiangxi Mo, Shiyi Cao, Joseph E Gonzalez, Ion Stoica, and Matei Zaharia. Optimizing llm queries in relational workloads. arXiv preprint arXiv:2403.05821, 2024

    2024 arXiv

  69. [77]

    Cost-intelligent data analytics in the cloud

    Huanchen Zhang, Yihao Liu, and Jiaqi Yan. Cost-intelligent data analytics in the cloud. arXiv preprint arXiv:2308.09569, 2023

    2023 arXiv

  70. [78]

    Cloud-native database systems at alibaba: Opportunities and challenges

    Feifei Li. Cloud-native database systems at alibaba: Opportunities and challenges. Proceedings of the VLDB Endowment, 12(12):2263–2272, 2019

    2019

  71. [79]

    Cypher: An evolving query language for property graphs

    Nadime Francis, Alastair Green, Paolo Guagliardo, Leonid Libkin, Tobias Lindaaker, Victor Marsault, Ste- fan Plantikow, Mats Rydberg, Petra Selmer, and Andr´es Taylor. Cypher: An evolving query language for property graphs. In Proceedings of the 2018 International Conference o...

    2018

  72. [80]

    Smola, and Zheng Zhang

    Minjie Wang, Lingfan Yu, Da Zheng, Quan Gan, Yu Gai, Zihao Ye, Mufei Li, Jinjing Zhou, Qi Huang, Chao Ma, Ziyue Huang, Qipeng Guo, Hao Zhang, Haibin Lin, Junbo Zhao, Jinyang Li, Alexander J. Smola, and Zheng Zhang. Deep graph library: Towards efficient and scalable deep learni...

    1909 arXiv

  73. [81]

    G- core: A core for future graph query languages

    Renzo Angles, Marcelo Arenas, Pablo Barcelo, Peter Boncz, George Fletcher, Claudio Gutierrez, Tobias Lindaaker, Marcus Paradies, Stefan Plantikow, Juan Sequeda, Oskar van Rest, and Hannes V oigt. G- core: A core for future graph query languages. In Proceedings of the 2018 Inte...

    2018

  74. [82]

    Reasoning over semantic-level graph for fact checking

    Wanjun Zhong, Jingjing Xu, Duyu Tang, Zenan Xu, Nan Duan, Ming Zhou, Jiahai Wang, and Jian Yin. Reasoning over semantic-level graph for fact checking. arXiv preprint arXiv:1909.03745, 2019

    1909 arXiv

  75. [83]

    Entity linking with a knowledge base: Issues, techniques, and solutions

    Wei Shen, Jianyong Wang, and Jiawei Han. Entity linking with a knowledge base: Issues, techniques, and solutions. IEEE Transactions on Knowledge and Data Engineering, 27(2):443–460, 2014

    2014

  76. [84]

    Eventground: Narrative reasoning by grounding to eventuality-centric knowledge graphs

    Cheng Jiayang, Lin Qiu, Chunkit Chan, Xin Liu, Yangqiu Song, and Zheng Zhang. Eventground: Narrative reasoning by grounding to eventuality-centric knowledge graphs. In Proceedings of the 2024 Joint Inter- national Conference on Computational Linguistics, Language Resources and...

    2024

  77. [85]

    Dense x retrieval: What retrieval granularity should we use? arXiv preprint arXiv:2312.06648, 2023

    Tong Chen, Hongwei Wang, Sihao Chen, Wenhao Yu, Kaixin Ma, Xinran Zhao, Hongming Zhang, and Dong Yu. Dense x retrieval: What retrieval granularity should we use? arXiv preprint arXiv:2312.06648, 2023

    2023 arXiv

  78. [86]

    Understanding inter-session intentions via complex logical reasoning, 2024

    Jiaxin Bai, Chen Luo, Zheng Li, Qingyu Yin, and Yangqiu Song. Understanding inter-session intentions via complex logical reasoning, 2024. 18

    2024

  79. [2022]

    arXiv:2110.14890 [cs]. 16

    arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.