Pith. sign in

REVIEW 4 major objections 4 minor 110 references

Panning for Gold: Expanding Domain-Specific Knowledge Graphs with General Knowledge

T0 review · 4 major / 4 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read The paper introduces domain-specific knowledge graph fusion (DKGF) and ExeFuse, a neuro-symbolic framework that treats facts as executable programs, claiming up to 9.5% relative gains in accuracy and F1 over the strongest baselines.

desk verdict A genuinely useful new task and benchmark suite, but the headline fusion gains are unverified: ExeFuse trains on gold fusion labels while the comparison methods do not, and the gold labels themselves are undisclosed. read the letter →

arxiv 2601.10485 v4 pith:X42CRZHA submitted 2026-01-15 cs.AI

Runhao Zhao , Weixin Zeng , Wentao Zhang , Chong Chen , Zhengpin Li , Xiang Zhao , Lei Chen This is my paper
classification cs.AI
keywords domain-specificknowledgegraphfusionenrichmentgeneral-to-domaintransferFact-as-Programneuro-symbolicreasoninggranularityalignmentdistantsupervisionbenchmarks
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

Domain-specific knowledge graphs, such as those used for political event forecasting, are often sparse compared with general knowledge graphs like Wikidata and YAGO. This paper defines a new task, domain-specific knowledge graph fusion (DKGF), which aims to mine relevant facts from general graphs and integrate them into the domain graph. To handle the two core difficulties — deciding whether a general fact is truly relevant to the domain, and adapting coarse-grained general facts to the fine-grained format the domain needs — the authors propose ExeFuse, a neuro-symbolic framework built on a 'Fact-as-Program' paradigm. Facts are encoded as program states, logic rules act as transition operators, and a target-space grounding step checks that the transformed fact lands in a valid region of the domain graph. Experiments on two new benchmarks report up to 9.5% relative improvement in accuracy and F1 over strong baselines, along with better transfer to unseen entities and lower cost than LLM-based methods.

What carries the argument

The central object is the Fact-as-Program paradigm, which treats a source fact as a program whose state lives in a continuous 'quasi-symbolic' space. The pipeline has three stages: (1) latent program induction, where a structure-aware neural predicate encodes the tuple (head, relation, tail); (2) neuro-symbolic execution, where a small set of affine transformations (logic rules) are weighted by a neural selector to move the state toward a domain-relevant region; and (3) target-space grounding, where an executability score computed by kernel density estimation over clustered domain fact prototypes verifies that the transformed state falls in the valid semantic region of the DKG. The affine ru

What would settle it

Re-annotate a random sample of the fused facts on DKGF(W-I) and DKGF(Y-I) with human judgments of relevance and granularity; if ExeFuse-selected facts do not show higher human-rated precision and recall than baseline-selected facts, the superior-fusion claim fails. Alternatively, change the entity-alignment step used to construct the gold labels; if the ranking of ExeFuse vs. baselines flips under different alignment choices, the benchmark is not stable.

Watch

Extended reading notes

Core claim

The central claim is that knowledge-graph fusion can be reformulated as an executable neuro-symbolic program: rather than matching facts by surface similarity, ExeFuse encodes each general fact into a continuous state via a structure-aware predicate, applies learned affine logic rules selected by a neural selector, and accepts the fact only if the transformed state grounds into the density of valid domain facts via a kernel-density executability score. This mechanism simultaneously resolves the ambiguity of domain relevance (by logical reachability) and calibrates granularity (by target-space grounding). The paper offers theoretical propositions for the expressiveness of the mapping and the

Load-bearing premise

The gold fused facts and training pairs were built by a distant-supervision and alignment process the paper does not fully specify; if those labels were derived from surface-level entity overlap or from the same embedding similarities the method is intended to surpass, the benchmark is circular and the reported improvements are artifacts.

Editorial extensions

If this is right

  • If the claim holds, DKGF becomes a standard, benchmarked task: future systems can be compared on the two new datasets with the 21 adapted baselines.
  • The success of a small non-LLM model over token-heavy LLM baselines suggests an efficient direction for knowledge enrichment where logical reachability suffices.
  • The transferability results on unseen entities imply that the learned logic rules are not memorizing entities but capturing transferable inference patterns, which is essential for real graphs that grow over time.
  • The ablation study shows each component (structure-aware encoding, rule selection, and grounding) contributes non-trivially, guiding what a future fusion model must include.

Reading between the lines

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

  • If the reported gains survive independent human annotation, the same neuro-symbolic execution recipe could transfer to other domain pairs (biomedical, legal, financial) where relevance and granularity gaps are equally severe.
  • The benchmark's validity hinges on the undisclosed distant-supervision process; an openly documented alignment and negative-sampling procedure would let the community verify that the gold labels are not accidentally encoding the same surface similarities the method is meant to beat.
  • The grounding mechanism's reliance on a fixed set of DKG prototypes may need to be adapted for dynamic domains (e.g., evolving political events) where the valid semantic region shifts over time.
  • Because the rule set is learned from training pairs, the framework's expressiveness depends on the quality and coverage of the distant supervision; with low-quality alignments, the 'logic' could become a learned rationalization of the label noise.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper introduces domain-specific knowledge graph fusion (DKGF), a task in which facts from a general knowledge graph (e.g., Wikidata, YAGO) are mined and integrated into a domain-specific knowledge graph (ICEWS). The authors propose ExeFuse, a neuro-symbolic framework built on a 'Fact-as-Program' paradigm: GKG facts are embedded into a continuous space, transformed by learned affine 'logic rules' selected by an attention mechanism, and then grounded via a KDE-based executability score against the DKG's fact embeddings. The paper also constructs two new benchmarks, DKGF(W-I) and DKGF(Y-I), and evaluates 21 adapted baselines plus ExeFuse on triple classification. The central claims are that ExeFuse outperforms all baselines by up to 9.5% relative ACC/F1, that its learned rules transfer to unseen entities, and that the benchmarks are the first standardized evaluation suite for DKGF.

Significance. If the claims are validated, the DKGF task formalization is a useful new direction and the two benchmarks would provide a common testbed for an under-explored problem. The breadth of the benchmark suite—21 baseline configurations spanning rule-based, embedding, GNN, LLM, EA, KGC, and RTE methods—is a genuine strength, as is the efficiency analysis and the availability of code and data (via an anonymous link). The ablation study also gives a reasonable internal picture of which components matter. However, the paper's central evidence is currently undermined by two problems: the gold fusion labels are constructed by an undisclosed process, and ExeFuse is the only method trained on those gold labels. Until these issues are addressed, the reported superiority of ExeFuse cannot be attributed to the proposed mechanism rather than to label access. The 'neuro-symbolic' interpretation of the learned affine rules is also not supported by the current formalism.

major comments (4)
  1. [§5.1.1, §4.5, Table 1] The construction of the gold fused facts F_fused and the training pairs D_pos is not described. §5.1.1 only says 'we align and fuse cross-domain knowledge,' and §4.5 states D_pos is 'constructed via distant supervision.' Table 1 reports 0% relation overlap between DKG and GKG, yet Eq. (1) requires each fused fact to use a relation r in R_d. The relation-mapping step is therefore essential and must be specified: how are GKG relations mapped to ICEWS relations? How are candidate GKG facts generated, how are positive pairs aligned, and how are negatives sampled? If the gold labels were derived from surface string overlap or from embedding similarity, the evaluation is circular and the reported gains would be artifacts. The authors should release or document the full construction protocol, including entity alignment, relation mapping, negative sampling, and inter-annotator or rule-based vali
  2. [§5.1.2 vs. §4.5, Table 2] The headline comparison is confounded by unequal access to gold fusion labels. ExeFuse is trained with a binary cross-entropy loss on D_pos and negative samples (Eq. 9) and a rule-regression loss on D_pos (Eq. 10), so it directly sees the gold fusion decision. The baselines, according to §5.1.2, are adapted 'without access to gold fusion labels': translational and GNN models are trained on the joint graph G_d ∪ G_g, EA methods use aligned entity pairs, and RTE methods 'operate purely on structured triples.' Thus the 'up to 9.5% relative improvement' in §5.2.1 may simply measure the value of having supervised fusion labels, not the value of neuro-symbolic execution. A matched-supervision ablation is needed—for example, training a simple classifier (or a TransE-style scorer) on the same D_pos/negative pairs and comparing. Without this, the central claim of superiority is not established.
  3. [§4.3.1, Eq. (4), Eq. (10)] The 'logic rules' L are not shown to be logical rules. Each rule is a learned affine transformation T_R(q) = w_R ⊙ q + b_R (Eq. 4), and the rule-consistency loss (Eq. 10) directly regresses the executed state q_hat(f_g) against the gold DKG embedding phi_dkg(f_d) for positive pairs. The 'logical reachability' is therefore a supervised learned mapping from source facts to target embeddings, not an independently motivated entailment mechanism. Proposition 1 is essentially the universal approximation theorem for mixtures of affine experts; it does not show that the learned transformations correspond to valid logical inference. This does not invalidate the method empirically, but it does undermine the 'Fact-as-Program' and 'neuro-symbolic' framing, and it weakens the transferability claim. Please provide evidence that the learned rules are interpretable or that they generalize to unseen logi
  4. [Table 3, §5.2.2] The transferability claim is overstated. For ExeFuse, the unseen-entity F1 is 0.565, which is substantially lower than the seen-entity F1 of 0.709, even though the unseen ACC is higher (0.731 vs. 0.672). While ExeFuse is indeed better than the listed baselines on unseen facts, the statement that it 'maintains consistently high F1 scores' is not supported in absolute terms. Moreover, no variance or statistical significance is reported for any result in Table 2 or Table 3, so it is unclear whether the margins over baselines are stable.
minor comments (4)
  1. [Throughout] The manuscript contains multiple typographical and formatting issues: 'challenges of challenges' in RQ1, 'DisMult-F' instead of 'DistMult-F' in Table 6, 'Graph-Memba' instead of 'Graph Mamba' in Table 2, and inconsistent spacing in the abstract. The page header 'Preprint. Work in progress.' should be removed for a journal submission.
  2. [References] There are duplicate references: [19] and [20] are both BERT (NAACL 2019), [49] and [50] are the same aspect-sentiment extraction paper, and [60] and [61] are the same 'Unifying LLMs and KGs' paper. Please deduplicate and verify all bibliography entries.
  3. [§5.2.1] The 'up to 9.5% relative improvement' number is quoted without showing the exact computation. Since readers may want to reproduce this from Table 2, state which two configurations are compared and give the formula (e.g., (0.690 − 0.630)/0.630 for F1).
  4. [§5.3, Table 4] The ablation variants are clearly named, but the 'w/o Rules L (Single Affine Projection)' row is described in the text as 'comparable to the translation-based baseline TransE-F (F1 0.630),' while the reported F1 is 0.612. This is a 1.8-point gap, not 'comparable' in a strict sense; please soften the wording or provide a significance test.

Circularity Check

0 steps flagged · score 0.0 of 10

No construction-level circularity; supervision asymmetry and undisclosed label provenance are evaluation-validity risks, not circular derivation.

full rationale

No circular step is demonstrable from the paper's own equations. ExeFuse's rule consistency loss (Eq. 10) regresses the executed state qhat_d(f_g) to the gold target embedding phi_dkg(f_d) for positive pairs, and the fusion network (Eq. 8) is trained with binary cross-entropy on D_pos and negatives (Eq. 9). This is supervised fitting followed by held-out evaluation, not a prediction that is equivalent to its training input by construction. The executability score S_exec (Eq. 7) is computed against KDE prototypes drawn from existing DKG facts F_d, not against the gold fused facts F_fused, so the validation signal is not the training target. The self-citations ([97], [104], [105]) appear only in related-work discussion of entity alignment and are not load-bearing for the DKGF claims. Two concerns raised by the reader/skeptic are real but are not circularity: (i) §5.1.1 and §4.5 leave the distant-supervision label-generation and alignment process unspecified ('we align and fuse cross-domain knowledge'; 'constructed via distant supervision'), and (ii) §5.1.2 states that RTE baselines operate 'without access to gold fusion labels' while ExeFuse is trained on D_pos, creating a supervision asymmetry that threatens the headline 9.5% comparison. These would become circular only if the undisclosed label-generation procedure were shown to use the same surface/embedding signal ExeFuse claims to surpass, for which the paper provides no evidence. Accordingly, no reduction-by-construction or load-bearing self-citation chain is established.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The central claim rests on (i) an undisclosed distant-supervision protocol that produces the gold labels, and (ii) the assumption that learned affine transformations in embedding space constitute 'logical execution' and that KDE proximity to existing DKG facts indicates valid granularity. The first is a domain assumption that cannot be checked from the paper; the second is an ad hoc framing presented as a theory.

free parameters (8)
  • Structure-aware encoder parameters W1, W2, b1 = not reported
    Learned on fusion labels (Eq. 2).
  • Affine logic-rule parameters w_R, b_R = not reported
    The 'logic rules' are learned affine transforms (Eq. 4); number of rules |L| not reported.
  • Rule-selector parameters W_sel and u_R = not reported
    Learned attention parameters (Eq. 5).
  • Fusion network Psi = not reported
    Learned classifier head (Eq. 8).
  • KDE temperature tau = not reported
    Hyperparameter of executability score (Eq. 7); no value given.
  • Prototype count K = not reported
    Number of K-means centroids for target space; no value or clustering details given.
  • Fusion threshold delta = not reported
    Decision threshold for p(fuse)>delta; value not reported.
  • Loss weight lambda = not reported
    Relative weight of rule consistency loss (Eq. 10).
assumptions (5)
  • standard math Universal approximation: mixture of affine experts with softmax weights approximates any continuous function on a compact set (Prop. 1).
    Standard mixture-of-experts result; does not specifically justify 'logical reachability'.
  • standard math KDE over prototypes approximates the target DKG data density (Prop. 2).
    Standard kernel density estimation assumption; requires appropriate bandwidth (tau) and representative prototypes.
  • domain assumption Distant supervision produced correct gold fusion labels for D_pos.
    Section 4.5 and 5.1.1 assert aligned pairs were constructed via distant supervision, but the protocol is not described; if labels are wrong, all results are invalid.
  • ad hoc to paper A GKG fact is fusable iff its transformed embedding lands in the high-density region of DKG fact embeddings.
    The executability score (Eq. 7) is used to decide fusion; this equates semantic embedding proximity to domain compatibility, an assumption not independently validated.
  • ad hoc to paper Affine transformations w_R * q + b_R correspond to valid logical entailment rules.
    The paper calls learned affine maps 'logic rules' (Section 4.3.1); no symbolic rule extraction or interpretability evidence is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Panning for Gold: Expanding Domain-Specific Knowledge Graphs with General Knowledge." pith.science (2026). https://pith.science/paper/X42CRZHA

@misc{pith2026260110485,
  author       = {Pith},
  title        = {Pith review of: Panning for Gold: Expanding Domain-Specific Knowledge Graphs with General Knowledge},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/X42CRZHA}},
  note         = {Machine review of arXiv:2601.10485}
}
read the original abstract

Domain-specific knowledge graphs (DKGs) are critical yet often suffer from limited coverage compared to General Knowledge Graphs (GKGs). Existing tasks to enrich DKGs rely primarily on extracting knowledge from external unstructured data or completing KGs through internal reasoning, but the scope and quality of such integration remain limited. This highlights a critical gap: little systematic exploration has been conducted on how comprehensive, high-quality GKGs can be effectively leveraged to supplement DKGs. To address this gap, we propose a new and practical task: domain-specific knowledge graph fusion (DKGF), which aims to mine and integrate relevant facts from general knowledge graphs into domain-specific knowledge graphs to enhance their completeness and utility. Unlike previous research, this new task faces two key challenges: (1) high ambiguity of domain relevance, i.e., difficulty in determining whether knowledge from a GKG is truly relevant to the target domain , and (2) cross-domain knowledge granularity misalignment, i.e., GKG facts are typically abstract and coarse-grained, whereas DKGs frequently require more contextualized, fine-grained representations aligned with particular domain scenarios. To address these, we present ExeFuse, a neuro-symbolic framework based on a novel Fact-as-Program paradigm. ExeFuse treats fusion as an executable process, utilizing neuro-symbolic execution to infer logical relevance beyond surface similarity and employing target space grounding to calibrate granularity. We construct new datasets to establish the first standardized evaluation suite for this task. Extensive experiments demonstrate that ExeFuse effectively overcomes domain barriers to achieve superior fusion performance.

Figures

Figures reproduced from arXiv: 2601.10485 by the authors.

Figure 1
Figure 1. Examples of domain-specific knowledge graph fusion. Figure 1(b) demonstrates the fusion of highly relevant entities [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The framework of ExeFuse for DKGF. the semantic focus to trace the logical path from the source fact to the target context. 4.3.2 Neural Rule Selector. To address the ambiguity of which rule applies, we employ a Neural Rule Selector that computes an attention distribution 𝑝(𝑅 | q 𝑔 ) over L: 𝑝(𝑅 | q 𝑔 ) = exp(u ⊤ 𝑅 tanh(W𝑠𝑒𝑙q 𝑔 )) Í 𝑅′ ∈ L exp(u ⊤ 𝑅′ tanh(W𝑠𝑒𝑙q 𝑔 )), (5) where W𝑠𝑒𝑙 ∈ R 𝑑×𝑑 projects the state into th… view at source ↗
Figure 3
Figure 3. The venn diagram of correct predictions on [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

110 extracted references · 19 canonical work pages

  1. [1]

    Bilal Abu-Salih. 2021. Domain-specific knowledge graphs: A survey.J. Netw. Comput. Appl.185 (2021), 103076. https://doi.org/10.1016/J.JNCA.2021.103076

    arXiv 2021

  2. [2]

    Kinjo, Núria Queralt-Rosinach, and Robert Hoehndorf

    Mona Alshahrani, Mohammad Asif Khan, Omar Maddouri, Akira R. Kinjo, Núria Queralt-Rosinach, and Robert Hoehndorf. 2017. Neuro-symbolic representation learning on biological knowledge graphs.Bioinform.33, 17 (2017), 2723–2730. https://doi.org/10.1093/BIOINFORMATICS/BTX275

    work page doi:10.1093/bioinformatics/btx275 2017

  3. [3]

    Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi

  4. [4]

    Ali Behrouz and Farnoosh Hashemi. 2024. Graph Mamba: Towards Learning on Graphs with State Space Models. InProceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD 2024, Barcelona, Spain, August 25-29, 2024, Ricardo Baeza-Yates and Francesco Bonchi (Eds.). ACM, 119–130. https://doi.org/10.1145/3637528.3672044

    arXiv 2024

  5. [5]

    Antoine Bordes, Nicolas Usunier, Alberto Garcia-Duran, Jason Weston, and Oksana Yakhnenko. 2013. Translating embeddings for modeling multi-relational data.Adv Neural Inf Process Syst26 (2013)

    2013

  6. [6]

    Li Cai, Xin Mao, Meirong Ma, Hao Yuan, Jianchao Zhu, and Man Lan. 2022. A Simple Temporal Information Matching Mechanism for Entity Alignment between Temporal Knowledge Graphs. InProceedings of the 29th International Conference on Computational Linguistics, COLING 2022, Gyeongju, Republic of Ko- rea, October 12-17, 2022. International Committee on Computa...

    2022

  7. [7]

    Hongtai Cao, Qihao Wang, Xiaodong Li, Matin Najafi, Kevin Chen-Chuan Chang, and Reynold Cheng. 2024. Large Subgraph Matching: A Comprehensive and Efficient Approach for Heterogeneous Graphs. In40th IEEE International Conference on Data Engineering, ICDE 2024, Utrecht, The Netherlands, May 13-16,

    2024

  8. [8]

    Nicola De Cao, Gautier Izacard, Sebastian Riedel, and Fabio Petroni. 2021. Au- toregressive Entity Retrieval. In9th International Conference on Learning Rep- resentations, ICLR 2021, Virtual Event, Austria, May 3-7, 2021. OpenReview.net. https://openreview.net/forum?id=5k8F6UU39V

    2021

Show all 110 references
  1. [10]

    Hanzhu Chen, Xu Shen, Qitan Lv, Jie Wang, Xiaoqi Ni, and Jieping Ye. 2024. SAC-KG: Exploiting Large Language Models as Skilled Automatic Constructors for Domain Knowledge Graph. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1...

    2024 doi

  2. [11]

    Chen Chen, Yufei Wang, Bing Li, and Kwok-Yan Lam. 2022. Knowledge Is Flat: A Seq2Seq Generative Framework for Various Knowledge Graph Completion. InCOLING. International Committee on Computational Linguistics, 4005–4017

    2022

  3. [12]

    Shengyuan Chen, Qinggang Zhang, Junnan Dong, Wen Hua, Qing Li, and Xiao Huang. 2024. Entity Alignment with Noisy Annotations from Large Language Models. InAdvances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing Systems 2024, Neu...

    2024

  4. [13]

    Muhao Chen, Yingtao Tian, Mohan Yang, and Carlo Zaniolo. 2017. Multilingual Knowledge Graph Embeddings for Cross-lingual Knowledge Alignment. In IJCAI. 1511–1517

    2017

  5. [14]

    Ziang Chen, Jialin Liu, Xiaohan Chen, Xinshang Wang, and Wotao Yin

  6. [15]

    Ziyang Chen, Dongfang Li, Xiang Zhao, Baotian Hu, and Min Zhang. 2024. Temporal Knowledge Question Answering via Abstract Reasoning Induction. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2024, Bangkok, ...

    2024 doi

  7. [16]

    Ahmed, Ryan A

    Kewei Cheng, Nesreen K. Ahmed, Ryan A. Rossi, Theodore Willke, and Yizhou Sun. 2024. Neural-Symbolic Methods for Knowledge Graph Reasoning: A Survey. ACM Trans. Knowl. Discov. Data18, 9, Article 225 (Nov. 2024), 44 pages. https: //doi.org/10.1145/3686806

    2024 doi

  8. [17]

    Rethinking the Capacity of Graph Neural Networks for Branch- ing Strategy. InAdvances in Neural Information Processing Systems 38: An- nual Conference on Neural Information Processing Systems 2024, NeurIPS 2024, Vancouver, BC, Canada, December 10 - 15, 2024, Amir Globersons, L...

    2024

  9. [18]

    Zhibin Chen, Yuting Wu, Yansong Feng, and Dongyan Zhao. 2022. Integrating Manifold Knowledge for Global Entity Linking with Heterogeneous Graphs. Data Intelligence4, 1 (2022), 20–40

    2022

  10. [19]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. InNAACL-HLT (1). Association for Computational Linguistics, 4171–4186

    2019

  11. [20]

    2024.Knowledge Graph Reasoning: A Neuro- Symbolic Perspective

    Kewei Cheng and Yizhou Sun. 2024.Knowledge Graph Reasoning: A Neuro- Symbolic Perspective. Springer Nature

    2024

  12. [21]

    Yew Ken Chia, Lidong Bing, Soujanya Poria, and Luo Si. 2022. RelationPrompt: Leveraging Prompts to Generate Synthetic Data for Zero-Shot Relation Triplet Extraction. InFindings of the ACL, Smaranda Muresan, Preslav Nakov, and Aline Villavicencio (Eds.). Association for Computa...

    2022 doi

  13. [22]

    Yifan Ding, Qingkai Zeng, and Tim Weninger. 2024. ChatEL: Entity Linking with Chatbots. InProceedings of the 2024 Joint International Conference on Computa- tional Linguistics, Language Resources and Evaluation, LREC/COLING 2024, 20-25 May, 2024, Torino, Italy, Nicoletta Calzo...

    2024

  14. [23]

    Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. InProceedings of the 2019 Conference of the North American Chapter of the Associ- ation for Computational Linguistics: Huma...

    2019

  15. [24]

    Linlin Ding, Mengjunyao Si, and Mo Li. 2025. RSGEA: Relationship Structure Line Graph for Semi-supervised Entity Alignment based on Edge Weight Ad- justment. InProceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGI...

    2025

  16. [25]

    Hossein Rajaby Faghihi, Quan Guo, Andrzej Uszok, Aliakbar Nafar, and Parisa Kordjamshidi. 2021. DomiKnowS: A Library for Integration of Symbolic Domain Knowledge in Deep Learning. InProceedings of the 2021 Conference on Empirical Methods in Natural Language Processing: System ...

    2021 doi

  17. [26]

    Yongkun Du, Zhineng Chen, Caiyan Jia, Xieping Gao, and Yu-Gang Jiang. 2025. Out of Length Text Recognition with Sub-String Matching. InAAAI-25, Spon- sored by the Association for the Advancement of Artificial Intelligence, February 25 - March 4, 2025, Philadelphia, PA, USA, To...

    2025 doi

  18. [27]

    Fredo Erxleben, Michael Günther, Markus Krötzsch, Julian Mendez, and Denny Vrandecic. 2014. Introducing Wikidata to the Linked Data Web. InThe Semantic Web - ISWC 2014 - 13th International Semantic Web Conference, Riva del Garda, Italy, October 19-23, 2014. Proceedings, Part I...

    2014

  19. [28]

    Jiaying Gong and Hoda Eldardiry. 2024. Prompt-based Zero-shot Relation Extraction with Semantic Knowledge Augmentation. InProceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation, LREC/COLING 2024, 20-25 May, 2024, ...

    2024

  20. [29]

    Congcong Ge, Xiaoze Liu, Lu Chen, Baihua Zheng, and Yunjun Gao. 2021. Make It Easy: An Effective End-to-End Entity Alignment Framework. InSIGIR ’21: The 44th International ACM SIGIR Conference on Research and Development in Information Retrieval, Virtual Event, Canada, July 11...

    2021

  21. [30]

    Sandra Geisler, Cinzia Cappiello, Irene Celino, David Chaves-Fraga, Anastasia Dimou, Ana Iglesias-Molina, Maurizio Lenzerini, Anisa Rula, Dylan Van Assche, Sascha Welten, and Maria-Esther Vidal. 2025. From Genesis to Maturity: Man- aging Knowledge Graph Ecosystems Through Life...

    2025

  22. [31]

    Silman, Greg Lersen, Fan Yang, Robert J

    Debashis Gupta, Aditi Golder, Luis Fernendez, Miles R. Silman, Greg Lersen, Fan Yang, Robert J. Plemmons, Sarra M. Alqahtani, and Victor Paúl Pauca

  23. [32]

    Arianna Graciotti, Nicolas Lazzari, Valentina Presutti, and Rocco Tripodi. 2025. Musical heritage historical entity linking.Artif. Intell. Rev.58, 5 (2025), 140. https://doi.org/10.1007/S10462-024-11102-9

    2025 doi

  24. [33]

    Qian Guo, Yi Guo, and Jin Zhao. 2024. KBPT: knowledge-based prompt tuning for zero-shot relation triplet extraction.PeerJ Comput. Sci.10 (2024), e2014. https://doi.org/10.7717/PEERJ-CS.2014

    2024 doi

  25. [34]

    Nan Huo, Reynold Cheng, Ben Kao, Wentao Ning, Nur Al Hasan Haldar, Xi- aodong Li, Jinyang Li, Mohammad Matin Najafi, Tian Li, and Ge Qu. 2024. ZeroEA: A Zero-Training Entity Alignment Framework via Pre-Trained Lan- guage Model.Proc. VLDB Endow.17, 7 (2024), 1765–1774. https://...

    2024

  26. [35]

    CoRRabs/2408.08972 (2024)

    ASGM-KG: Unveiling Alluvial Gold Mining Through Knowledge Graphs. CoRRabs/2408.08972 (2024). https://doi.org/10.48550/ARXIV.2408.08972 arXiv:2408.08972

    2024 doi arXiv

  27. [36]

    Phillips, John H

    Keywan Hassani-Pak, Ajit Singh, Marco Brandizi, Joseph Hearnshaw, Sandeep Amberkar, Andrew L. Phillips, John H. Doonan, and Chris Rawlings. 2020. KnetMiner: a comprehensive approach for support- ing evidence-based gene discovery and complex trait analysis across species.bioRxi...

    2020 doi

  28. [37]

    Xuming Hu, Junzhe Chen, Shiao Meng, Lijie Wen, and Philip S. Yu. 2023. SelfLRE: Self-refining Representation Learning for Low-resource Relation Extraction. InProceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR ...

    2023

  29. [38]

    Di Jin, Bunyamin Sisman, Hao Wei, Xin Luna Dong, and Danai Koutra. 2021. Deep Transfer Learning for Multi-source Entity Linkage via Domain Adaptation. Proc. VLDB Endow.15, 3 (2021), 465–477. https://doi.org/10.14778/3494124. 3494131

    2021 doi

  30. [39]

    Shaoxiong Ji, Shirui Pan, Erik Cambria, Pekka Marttinen, and Philip S. Yu. 2022. A Survey on Knowledge Graphs: Representation, Acquisition, and Applications. IEEE Trans. Neural Networks Learn. Syst.33, 2 (2022), 494–514. https://doi.org/ 10.1109/TNNLS.2021.3070843

    2022

  31. [40]

    Xuhui Jiang, Yinghan Shen, Zhichao Shi, Chengjin Xu, Wei Li, Zixuan Li, Jian Guo, Huawei Shen, and Yuanzhuo Wang. 2024. Unlocking the Power of Large Language Models for Entity Alignment. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics ...

    2024

  32. [41]

    Arijit Khan, Tianxing Wu, and Xi Chen. 2024. LLM+KG: Data Management Opportunities in Unifying Large Language Models + Knowledge Graphs. In Proceedings of Workshops at the 50th International Conference on Very Large Data Bases, VLDB 2024, Guangzhou, China, August 26-30, 2024. ...

    2024

  33. [42]

    Xuhui Jiang, Chengjin Xu, Yinghan Shen, Yuanzhuo Wang, Fenglong Su, Zhichao Shi, Fei Sun, Zixuan Li, Jian Guo, and Huawei Shen. 2024. Toward Practical Entity Alignment Method Design: Insights from New Highly Heterogeneous Knowledge Graph Datasets. InProceedings of the ACM on W...

    2024

  34. [43]

    Mitchell

    Bosung Kim, Hayate Iso, Nikita Bhutani, Estevam Hruschka, Ndapa Nakashole, and Tom M. Mitchell. 2023. Zero-shot Triplet Extraction by Template Infilling. Association for Computational Linguistics, 272–284. https://doi.org/10.18653/ V1/2023.IJCNLP-MAIN.18

    2023

  35. [44]

    2019.Domain-Specific Knowledge Graph Construction

    Mayank Kejriwal. 2019.Domain-Specific Knowledge Graph Construction. Springer. https://doi.org/10.1007/978-3-030-12375-8

    2019 doi

  36. [45]

    Arijit Khan, Xiangyu Ke, and Yinghui Wu. 2025. Graph Data Management and Graph Machine Learning: Synergies and Opportunities.SIGMOD Rec.54, 2 (2025), 28–42. https://doi.org/10.1145/3749116.3749122

    2025

  37. [46]

    Jin Li, Zezhong Ding, and Xike Xie. 2025. DuetGraph: Coarse-to-Fine Knowledge Graph Reasoning with Dual-Pathway Global-Local Fusion.CoRRabs/2507.11229 (2025). https://doi.org/10.48550/ARXIV.2507.11229 arXiv:2507.11229

    2025 doi

  38. [47]

    Bosung Kim, Taesuk Hong, Youngjoong Ko, and Jungyun Seo. 2020. Multi-Task Learning for Knowledge Graph Completion with Pre-trained Language Models. InCOLING. International Committee on Computational Linguistics, 1737–1743

    2020

  39. [48]

    Ying Li, Vitalii Zakhozhyi, Daniel Zhu, and Luis J. Salazar. 2020. Domain Specific Knowledge Graphs as a Service to the Public: Powering Social-Impact Funding in the US. InKDD ’20: The 26th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, Virtual Event, CA, USA, A...

    2020

  40. [49]

    Haridimos Kondylakis, Stefania Dumbrava, Matteo Lissandrini, Nikolay Yakovets, Angela Bonifati, Vasilis Efthymiou, George Fletcher, Dimitris Plex- ousakis, Riccardo Tommasini, Georgia Troullinou, and Elisjana Ymeralli. 2025. Property Graph Standards: State of the Art & Open Ch...

    2025

  41. [50]

    Vignesh Kothapalli, Hamed Firooz, and Maziar Sanjabi. 2025. CoT-ICL Lab: A Synthetic Framework for Studying Chain-of-Thought Learning from In-Context Demonstrations. InProceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Paper...

    2025

  42. [51]

    Zehan Li, Fu Zhang, Kailun Lyu, Jingwei Cheng, and Tianyue Peng. 2025. Re- Cent: A Relation-Centric Framework for Joint Zero-Shot Relation Triplet Ex- traction. InProceedings of the 31st International Conference on Computational Linguistics, COLING 2025, Abu Dhabi, UAE, Januar...

    2025

  43. [52]

    Wanli Li, Tieyun Qian, Yi Song, Zeyu Zhang, Jiawei Li, Zhuang Chen, and Lixin Zou. 2025. Generative Meta-Learning for Zero-Shot Relation Triplet Extraction. InProceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR...

    2025 doi

  44. [53]

    Xueyuan Lin, Haihong E, Chengjin Xu, Gengxian Zhou, Haoran Luo, Tianyi Hu, Fenglong Su, Ningyuan Li, and Mingzhi Sun. 2023. TFLEX: Tem- poral Feature-Logic Embedding Framework for Complex Reasoning over Temporal Knowledge Graph. InAdvances in Neural Information Process- ing Sy...

    2023

  45. [55]

    You Li, Xupeng Zeng, Yixiao Zeng, and Yuming Lin. 2024. Enhanced Packed Marker with Entity Information for Aspect Sentiment Triplet Extraction. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2024, Washi...

    2024 doi

  46. [56]

    Xin Lv, Yankai Lin, Yixin Cao, Lei Hou, Juanzi Li, Zhiyuan Liu, Peng Li, and Jie Zhou. 2022. Do Pre-trained Models Benefit Knowledge Graph Completion? A Reliable Evaluation and a Reasonable Approach. InACL (Findings). Association for Computational Linguistics, 3570–3581

    2022

  47. [57]

    Hongtao Lin, Jun Yan, Meng Qu, and Xiang Ren. 2019. Learning Dual Re- trieval Module for Semi-supervised Relation Extraction. InThe World Wide Web Conference, WWW. 1073–1083

    2019

  48. [58]

    Xin Mao, Wenting Wang, Yuanbin Wu, and Man Lan. 2021. Boosting the Speed of Entity Alignment 10×: Dual Attention Matching Network with Normalized Hard Sample Mining. InWWW ’21: The Web Conference 2021, Virtual Event / Ljubljana, Slovenia, April 19-23, 2021. ACM / IW3C2, 821–832

    2021

  49. [59]

    Lihui Liu, Zihao Wang, and Hanghang Tong. 2025. Neural-Symbolic Reasoning over Knowledge Graphs: A Survey from a Query Perspective.SIGKDD Explor. 27, 1 (2025), 124–136. https://doi.org/10.1145/3748239.3748249

    2025

  50. [60]

    Xiaoze Liu, Junyang Wu, Tianyi Li, Lu Chen, and Yunjun Gao. 2023. Unsuper- vised Entity Alignment for Temporal Knowledge Graphs. InProceedings of the ACM Web Conference 2023, WWW 2023, Austin, TX, USA, 30 April 2023 - 4 May 2023, Ying Ding, Jie Tang, Juan F. Sequeda, Lora Aroy...

    2023

  51. [61]

    Shirui Pan, Linhao Luo, Yufei Wang, Chen Chen, Jiapu Wang, and Xindong Wu

  52. [62]

    Xin Lv, Yankai Lin, Yixin Cao, Lei Hou, Juanzi Li, Zhiyuan Liu, Peng Li, and Jie Zhou. 2022. Do Pre-trained Models Benefit Knowledge Graph Completion? A Reliable Evaluation and a Reasonable Approach. InFindings of the Association for Computational Linguistics: ACL 2022, Dublin...

    2022 doi

  53. [63]

    Derek Paulsen, Yash Govind, and AnHai Doan. 2023. Sparkly: A Simple yet Surprisingly Strong TF/IDF Blocker for Entity Matching.Proc. VLDB Endow.16, 6 (2023), 1507–1519. https://doi.org/10.14778/3583140.3583163

    2023

  54. [64]

    Jixuan Nie, Xia Hou, Wenfeng Song, Xuan Wang, Xingliang Jin, Xinyu Zhang, ShuoZhe Zhang, and Jiaqi Shi. 2024. Knowledge Graph Efficient Construction: Embedding Chain-of-Thought into LLMs. InProceedings of Workshops at the 50th International Conference on Very Large Data Bases,...

    2024

  55. [67]

    Taoyu Su, Jiawei Sheng, Duohe Ma, Xiaodong Li, Juwei Yue, Mengxiao Song, Yingkai Tang, and Tingwen Liu. 2025. Mitigating Modality Bias in Multi-modal Entity Alignment from a Causal Perspective. InProceedings of the 48th Inter- national ACM SIGIR Conference on Research and Deve...

    2025

  56. [68]

    IEEE Trans

    Unifying Large Language Models and Knowledge Graphs: A Roadmap. IEEE Trans. Knowl. Data Eng.36, 7 (2024), 3580–3599. https://doi.org/10.1109/ TKDE.2024.3352100

    2024

  57. [69]

    Heiko Paulheim. 2017. Knowledge graph refinement: A survey of approaches and evaluation methods.Semantic Web8, 3 (2017), 489–508. https://doi.org/10. 3233/SW-160218

    2017

  58. [70]

    Zequn Sun, Wei Hu, Qingheng Zhang, and Yuzhong Qu. 2018. Bootstrapping Entity Alignment with Knowledge Graph Embedding. InProceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, IJCAI 2018, July 13-19, 2018, Stockholm, Sweden. ijcai.org, 4396–4402

    2018

  59. [71]

    Larsen, and Manfred Jaeger

    Raffaele Pojer, Andrea Passerini, Kim G. Larsen, and Manfred Jaeger. 2025. A Neuro-Symbolic Approach for Probabilistic Reasoning on Graph Data. CoRRabs/2507.21873 (2025). https://doi.org/10.48550/ARXIV.2507.21873 arXiv:2507.21873

    2025 doi arXiv

  60. [72]

    Apoorv Saxena, Adrian Kochsiek, and Rainer Gemulla. 2022. Sequence-to- Sequence Knowledge Graph Completion and Question Answering. InACL (1). Association for Computational Linguistics, 2814–2828

    2022

  61. [73]

    Giuseppe Spillo, Cataldo Musto, Marco de Gemmis, Pasquale Lops, and Giovanni Semeraro. 2024. Recommender systems based on neuro-symbolic knowledge graph embeddings encoding first-order logic rules.User Model. User Adapt. Interact.34, 5 (2024), 2039–2083. https://doi.org/10.100...

    2024 doi

  62. [74]

    van Hulst, Faegheh Hasibi, Koen Dercksen, Krisztian Balog, and Arjen P

    Johannes M. van Hulst, Faegheh Hasibi, Koen Dercksen, Krisztian Balog, and Arjen P. de Vries. 2020. REL: An Entity Linker Standing on the Shoulders of Giants. InProceedings of the 43rd International ACM SIGIR conference on research and development in Information Retrieval, SIG...

    2020

  63. [75]

    Suchanek, Gjergji Kasneci, and Gerhard Weikum

    Fabian M. Suchanek, Gjergji Kasneci, and Gerhard Weikum. 2007. Yago: a core of semantic knowledge. InProceedings of the 16th International Conference on World Wide Web, WWW 2007, Banff, Alberta, Canada, May 8-12, 2007. ACM, 697–706

    2007

  64. [76]

    Zhiqing Sun, Zhi-Hong Deng, Jian-Yun Nie, and Jian Tang. 2019. RotatE: Knowl- edge Graph Embedding by Relational Rotation in Complex Space. InICLR (Poster). OpenReview.net

    2019

  65. [77]

    Jue Wang and Wei Lu. 2020. Two are Better than One: Joint Entity and Relation Extraction with Table-Sequence Encoders. InEMNLP, Bonnie Webber, Trevor Cohn, Yulan He, and Yang Liu (Eds.). Association for Computational Linguistics, 1706–1721. https://doi.org/10.18653/v1/2020.emn...

    2020 doi

  66. [78]

    Xiaobin Tang, Jing Zhang, Bo Chen, Yang Yang, Hong Chen, and Cuiping Li

  67. [79]

    Yaxiong Wang, Lianwei Wu, Lechao Cheng, Zhun Zhong, Yujiao Wu, and Meng Wang. 2025. Beyond General Alignment: Fine-Grained Entity-Centric Image- Text Matching with Multimodal Attentive Experts. InProceedings of the 48th International ACM SIGIR Conference on Research and Develo...

    2025

  68. [80]

    Xiaobin Tian, Zequn Sun, and Wei Hu. 2024. Generating Explanations to Under- stand and Repair Embedding-Based Entity Alignment. In40th IEEE International Conference on Data Engineering, ICDE 2024, Utrecht, The Netherlands, May 13-16,

    2024

  69. [82]

    Théo Trouillon, Johannes Welbl, Sebastian Riedel, Éric Gaussier, and Guillaume Bouchard. 2016. Complex Embeddings for Simple Link Prediction. InICML (JMLR Workshop and Conference Proceedings, Vol. 48). JMLR.org, 2071–2080

    2016

  70. [83]

    Junyang Wu, Tianyi Li, Lu Chen, Yunjun Gao, and Ziheng Wei. 2023. SEA: A Scalable Entity Alignment System. InProceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR 2023, Taipei, Taiwan, July 23-27, 2023, Hsin-Hsi ...

    2023

  71. [84]

    Bo Wang, Tao Shen, Guodong Long, Tianyi Zhou, Ying Wang, and Yi Chang

  72. [85]

    Chengjin Xu, Fenglong Su, and Jens Lehmann. 2021. Time-aware Graph Neu- ral Network for Entity Alignment between Temporal Knowledge Graphs. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, EMNLP 2021, Virtual Event / Punta Cana, Domini...

    2021

  73. [86]

    Chenxu Wang, Zhenhao Huang, Yue Wan, Junyu Wei, Junzhou Zhao, and Pinghui Wang. 2023. FuAlign: Cross-lingual entity alignment via multi-view representation learning of fused knowledge graphs.Inform. Fusion89 (Jan. 2023), 41–52. https://doi.org/10.1016/j.inffus.2022.08.002

    2023 doi

  74. [87]

    Linyao Yang, Hongyang Chen, Xiao Wang, Jing Yang, Fei-Yue Wang, and Han Liu. 2024. Two Heads Are Better Than One: Integrating Knowledge from Knowledge Graphs and Large Language Models for Entity Alignment.ArXiv abs/2401.16960 (2024). https://api.semanticscholar.org/CorpusID:267321024

    2024 arXiv

  75. [88]

    Le, Ed H

    Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc V. Le, Ed H. Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. 2023. Self-Consistency Improves Chain of Thought Reasoning in Language Models. InThe Eleventh International Conference on Learning Representations, ICLR 2023, Ki...

    2023

  76. [89]

    Rui Yang, Boming Yang, Xinjie Zhao, Fan Gao, Aosong Feng, Sixun Ouyang, Moritz Blum, Tianwei She, Yuang Jiang, Freddy Lécué, Jinghui Lu, and Irene Li. 2025. Graphusion: A RAG Framework for Scientific Knowledge Graph Construction with a Global Perspective. InCompanion Proceedin...

    2025

  77. [90]

    Zhichun Wang, Qingsong Lv, Xiaohan Lan, and Yu Zhang. 2018. Cross-lingual Knowledge Graph Alignment via Graph Convolutional Networks. InProceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Brussels, Belgium, October 31 - November 4, 2018. Ass...

    2018

  78. [91]

    Chi, Quoc V

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Brian Ichter, Fei Xia, Ed H. Chi, Quoc V. Le, and Denny Zhou. 2022. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. InAdvances in Neural Information Processing Systems 35: Annual Conference on Neura...

    2022

  79. [92]

    Yuyang Wei, Wei Chen, Xiaofang Zhang, Pengpeng Zhao, Jianfeng Qu, and Lei Zhao. 2024. Multi-Modal Siamese Network for Few-Shot Knowledge Graph Completion. In40th IEEE International Conference on Data Engineering, ICDE 2024, Utrecht, The Netherlands, May 13-16, 2024. IEEE, 719–...

    2024

  80. [93]

    Deming Ye, Yankai Lin, Peng Li, and Maosong Sun. 2022. Packed Levitated Marker for Entity and Relation Extraction. InProceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2022, Dublin, Ireland, May 22-27, 2022, Sma...

    2022

  81. [94]

    Ledell Wu, Fabio Petroni, Martin Josifoski, Sebastian Riedel, and Luke Zettle- moyer. 2020. Scalable Zero-shot Entity Linking with Dense Entity Retrieval. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020, Online, November 16...

    2020 doi

  82. [95]

    Donghan Yu, Chenguang Zhu, Yuwei Fang, Wenhao Yu, Shuohang Wang, Yi- chong Xu, Xiang Ren, Yiming Yang, and Michael Zeng. 2022. KG-FiD: In- fusing Knowledge Graph in Fusion-in-Decoder for Open-Domain Question Answering. InProceedings of the 60th Annual Meeting of the Associatio...

    2022

  83. [96]

    Bishan Yang, Wen-tau Yih, Xiaodong He, Jianfeng Gao, and Li Deng. 2015. Embedding Entities and Relations for Learning and Inference in Knowledge Bases. InICLR (Poster)

    2015

  84. [97]

    Weixin Zeng, Jie Zhou, and Xiang Zhao. 2024. Benchmarking Challenges for Temporal Knowledge Graph Alignment.Proceedings of the 33rd ACM International Conference on Information and Knowledge Management(2024). https://api.semanticscholar.org/CorpusID:273501043

    2024

  85. [98]

    Mengzheng Yang, Yanfei Ren, David Osei Opoku, Ruochang Li, Peng Ren, and Chunxiao Xing. 2025. DSRAG: A Domain-Specific Retrieval Framework Based on Document-derived Multimodal Knowledge Graph.CoRRabs/2509.10467 (2025). https://doi.org/10.48550/ARXIV.2509.10467 arXiv:2509.10467

    2025 doi arXiv

  86. [99]

    Peiyan Zhang, Yuchen Yan, Xi Zhang, Chaozhuo Li, Senzhang Wang, Feiran Huang, and Sunghun Kim. 2024. TransGNN: Harnessing the Collaborative Power of Transformers and Graph Neural Networks for Recommender Systems. InProceedings of the 47th International ACM SIGIR Conference on ...

    2024

  87. [101]

    Liang Yao, Chengsheng Mao, and Yuan Luo. 2019. KG-BERT: BERT for Knowledge Graph Completion.CoRRabs/1909.03193 (2019). arXiv:1909.03193 http://arxiv.org/abs/1909.03193

    2019 arXiv

  88. [102]

    Liang Yao, Jiazhen Peng, Chengsheng Mao, and Yuan Luo. 2025. Exploring Large Language Models for Knowledge Graph Completion. In2025 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2025, Hyderabad, India, April 6-11, 2025. IEEE, 1–5. https://doi...

    2025

  89. [103]

    Zhanqiu Zhang, Jie Wang, Jieping Ye, and Feng Wu. 2022. Rethinking Graph Convolutional Networks in Knowledge Graph Completion. InWWW ’22: The ACM Web Conference 2022, Virtual Event, Lyon, France, April 25 - 29, 2022, Frédérique Laforest, Raphaël Troncy, Elena Simperl, Deepak A...

    2022

  90. [104]

    Runhao Zhao, Weixin Zeng, Jiuyang Tang, Yawen Li, Guanhua Ye, Junping Du, and Xiang Zhao. 2025. Towards Unsupervised Entity Alignment for Highly Heterogeneous Knowledge Graphs . In2025 IEEE 41st International Conference on Data Engineering (ICDE). IEEE Computer Society, Los Al...

    2025

  91. [105]

    Hongbin Ye, Ningyu Zhang, Hui Chen, and Huajun Chen. 2022. Generative Knowledge Graph Construction: A Review. InEMNLP. Association for Compu- tational Linguistics, 1–17

    2022

  92. [107]

    Weixin Zeng, Xiang Zhao, Wei Wang, Jiuyang Tang, and Zhen Tan. 2020. Degree- Aware Alignment for Entities in Tail. InProceedings of the 43rd International ACM SIGIR conference on research and development in Information Retrieval, SIGIR 2020, Virtual Event, China, July 25-30, 2...

    2020

  93. [109]

    Ningyu Zhang, Zekun Xi, Yujie Luo, Peng Wang, Bozhong Tian, Yunzhi Yao, Jintian Zhang, Shumin Deng, Mengshu Sun, Lei Liang, Zhiqiang Zhang, Xiaowei Zhu, Jun Zhou, and Huajun Chen. 2024. OneEdit: A Neural-Symbolic Collab- oratively Knowledge Editing System. InProceedings of Wor...

    2024

  94. [111]

    Yutong Zhang, Lixing Chen, Shenghong Li, Nan Cao, Yang Shi, Jiaxin Ding, Zhe Qu, Pan Zhou, and Yang Bai. 2025. Way to Specialist: Closing Loop Between Specialized LLM and Evolving Domain Knowledge Graph. InProceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery an...

    2025

  95. [112]

    Yichi Zhang, Zhuo Chen, Yin Fang, Yanxi Lu, Fangming Li, Wen Zhang, and Huajun Chen. 2024. Knowledgeable Preference Alignment for LLMs in Domain- specific Question Answering. InFindings of the Association for Computational Linguistics, ACL 2024, Bangkok, Thailand and virtual m...

    2024 doi

  96. [113]

    Yichi Zhang, Zhuo Chen, Lingbing Guo, Yajing Xu, Wen Zhang, and Huajun Chen. 2024. Making Large Language Models Perform Better in Knowledge Graph Completion. InProceedings of the 32nd ACM International Conference on Multimedia, MM 2024, Melbourne, VIC, Australia, 28 October 20...

    2024

  97. [116]

    Runhao Zhao, Weixin Zeng, Wentao Zhang, Xiang Zhao, Jiuyang Tang, and Lei Chen. 2025. Towards Temporal Knowledge Graph Alignment in the Wild.CoRRabs/2507.14475 (2025). https://doi.org/10.48550/ARXIV.2507.14475 arXiv:2507.14475

    2025 doi arXiv

  98. [2020]

    BERT}-{INT: A {BERT}-based interaction model for knowledge graph alignment.interactions100 (2020), e1

    2020

  99. [2021]

    Structure-Augmented Text Representation Learning for Efficient Knowl- edge Graph Completion. InWWW. ACM / IW3C2, 1737–1748

  100. [2024]

    InThe Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024

    Self-RAG: Learning to Retrieve, Generate, and Critique through Self- Reflection. InThe Twelfth International Conference on Learning Representations, ICLR 2024, Vienna, Austria, May 7-11, 2024. OpenReview.net. https://openreview. net/forum?id=hSyW5go0v8

    2024

  101. [4917]

    https://doi.org/10.18653/v1/2022.acl-long.337

    2022 doi

  102. [7583]

    https://aclanthology.org/2024.acl-long.408

    2024

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.