Elamur Research⌕
☰
home / papers / 2026 / Structure-Aware Zero-Shot Relational Learning for Knowledge Graphs without External Knowledge

Structure-Aware Zero-Shot Relational Learning for Knowledge Graphs without External Knowledge

Paper · 2026 · Kuan Xu, Baoxin Zhang, Shuyue Fan et al.

Summary

Zero-shot relational learning (ZRL) asks a model to complete knowledge graph (KG) triples for relations that have no training examples. Earlier methods connect new relations to known ones through side information such as textual descriptions, ontological schemas or expert-assigned relevance scores. That information is expensive to produce and often missing. Xu et al. propose SAZRL, which uses only the graph's own topology. It assumes that relations with related meanings tend to link the same entities.

For a query with an unseen relation, the Conditional Query Graphs (CQG) module samples a two-hop subgraph around the query entities. It then builds a graph whose nodes are relations and whose weighted edges reflect how many entities they share. The Adaptive Relation Updating (ARU) module creates an embedding for the new relation by attention-weighted aggregation of its seen neighbours. Training maximizes the likelihood of observed triples under a relation–entity interaction score.

On NELL-ZS, Wiki-ZS and a newly built FB15K-ZS, SAZRL outperforms state-of-the-art ZRL baselines. This shows that annotation-free ZRL is practical.

Key points

  • SAZRL improves MRR by 10.66% on NELL-ZS and 4.33% on Wiki-ZS over the strongest prior ZRL methods, without text, ontologies or expert labels.
  • FB15K-ZS is a new benchmark derived from FB15K-237 that deliberately contains no external knowledge.
  • During training, seen triples are re-split dynamically into a background graph and a simulated "unseen" set. At test time only one query pair is used, to prevent leakage. Entities are assumed closed-world.
  • A sparsity study fixed shared entities at one and kept n ∈ {0, 1, 3, 10, 20} neighbour relations. MRR grew and then levelled off as neighbours increased, and very few neighbours hurt performance.
  • A NELL-ZS case study showed the top seen relations retrieved for "Wife of": Has spouse, Mother of person and Has family member.
  • The authors list two limitations: similar relations without shared entities are missed, and KG noise weakens structural inference.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Beijing Jiaotong UniversityBeijing · China28 works · 1,754 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 works

🧭 Topics 1Research subjects and languages the work is about.Knowledge graphs & ontologies · 211 works

Cite

Cite as Kuan Xu, Baoxin Zhang, Shuyue Fan, Ming Chen, Zhipeng Ke, Jian Yu and Xuezhong Zhou. 2026. Structure-Aware Zero-Shot Relational Learning for Knowledge Graphs without External Knowledge. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.941

@inproceedings{xu2026structure,
  title = {{Structure-Aware Zero-Shot Relational Learning for Knowledge Graphs without External Knowledge}},
  author = {Kuan Xu, Baoxin Zhang, Shuyue Fan, Ming Chen, Zhipeng Ke, Jian Yu and Xuezhong Zhou},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.941},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.941}
}

References

33 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Shaked Brody, Uri Alon, and Eran Yahav. 2021. How attentive are graph attention networks? arXiv preprint arXiv:2105.14491. — [arXiv:2105.14491](https://arxiv.org/abs/2105.14491) find
  2. Dan Busbridge, Dane Sherburn, Pietro Cavallo, and Nils Y Hammerla. 2019. Relational graph attention networks. arXiv preprint arXiv:1904.05811. — [arXiv:1904.05811](https://arxiv.org/abs/1904.05811) find
  3. Mingyang Chen, Wen Zhang, Yuxia Geng, Zezhong Xu, Jeff Z Pan, and Huajun Chen. 2023. Generalizing to unseen elements: A survey on knowledge extrapolation for knowledge graphs. arXiv preprint arXiv:2302.01859. — [arXiv:2302.01859](https://arxiv.org/abs/2302.01859) find
  4. Zhuo Chen, Yichi Zhang, Yin Fang, Yuxia Geng, Lingbing Guo, Xiang Chen, Qian Li, Wen Zhang, Jiaoyan Chen, Yushan Zhu, and 1 others. 2024. Knowledge graphs meet multi-modal learning: A comprehensive survey. arXiv preprint arXiv:2402.05391. — [arXiv:2402.05391](https://arxiv.org/abs/2402.05391) find
  5. Ziqiang Cui, Yunpeng Weng, Xing Tang, Fuyuan Lyu, Dugang Liu, Xiuqiang He, and Chen Ma. 2025. Comprehending knowledge graphs with large language models for recommender systems. In Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 1… find
  6. Ling Ding, Lei Huang, Zhizhi Yu, Di Jin, and Dongxiao He. 2025. Towards global-topology relation graph for inductive knowledge graph completion. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 11581–11589. find
  7. Zhijun Dong, Likang Wu, Kai Zhang, Ye Liu, Yanghai Zhang, Zhi Li, Hongke Zhao, and Enhong Chen. 2024. Fzr: Enhancing knowledge transfer via shared factors composition in zero-shot relational learning. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, pa… find
  8. John Duchi, Elad Hazan, and Yoram Singer. 2011. Adaptive subgradient methods for online learning and stochastic optimization. Journal of machine learning research, 12(7). find
  9. Zhiyi Fang, Hang Yu, Changhua Xu, Zhuofeng Li, Jie Ying, and Shaorong Xie. 2025. Contrastive zero-shot relational learning for knowledge graph completion. Knowledge-Based Systems, page 113425. find
  10. Mikhail Galkin, Xinyu Yuan, Hesham Mostafa, Jian Tang, and Zhaocheng Zhu. 2024. Towards foundation models for knowledge graph reasoning. In The Twelfth International Conference on Learning Representations. find
  11. Xiou Ge, Yun Cheng Wang, Bin Wang, and C-C Jay Kuo. 2023. Compounding geometric operations for knowledge graph completion. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 6947–6965. find
  12. Yuxia Geng, Jiaoyan Chen, Zhuo Chen, Jeff Z Pan, Zhiquan Ye, Zonggang Yuan, Yantao Jia, and Huajun Chen. 2021. Ontozsl: Ontology-enhanced zero-shot learning. In Proceedings of the web conference 2021, pages 3325–3336. find
  13. Yuxia Geng, Jiaoyan Chen, Wen Zhang, Yajing Xu, Zhuo Chen, Jeff Z. Pan, Yufeng Huang, Feiyu Xiong, and Huajun Chen. 2022. Disentangled ontology embedding for zero-shot learning. In Proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining, pages 443–453. find
  14. Xavier Glorot and Yoshua Bengio. 2010. Understanding the difficulty of training deep feedforward neural networks. In Proceedings of the thirteenth international conference on artificial intelligence and statistics, pages 249–256. JMLR Workshop and Conference Proceedings. find
  15. Xavier Glorot, Antoine Bordes, and Yoshua Bengio. 2011. Deep sparse rectifier neural networks. In Proceedings of the fourteenth international conference on artificial intelligence and statistics, pages 315– 323. JMLR Workshop and Conference Proceedings. find
  16. Christian Meilicke, Melisachew Wudage Chekol, Patrick Betz, Manuel Fink, and Heiner Stuck-eschmidt. 2024. Anytime bottom-up rule learning for large-scale knowledge graph completion. The VLDB Journal, 33(1):131–161. find
  17. Pengda Qin, Xin Wang, Wenhu Chen, Chunyun Zhang, Weiran Xu, and William Yang Wang. 2020. Generative adversarial zero-shot relational learning for knowledge graphs. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 8673–8680. find
  18. Andrea Rossi, Denilson Barbosa, Donatella Firmani, Antonio Matinata, and Paolo Merialdo. 2021. Knowledge graph embedding for link prediction: A comparative analysis. ACM Transactions on Knowledge Discovery from Data (TKDD), 15(2):1–49. find
  19. Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. 2020. Generative adversarial networks. Communications of the ACM, 63(11):139–144. find
  20. Kristina Toutanova and Danqi Chen. 2015. Observed versus latent features for knowledge base and text inference. In Proceedings of the 3rd workshop on continuous vector space models and their compositionality, pages 57–66.
  21. Thomas N Kipf and Max Welling. 2016. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907.
  22. Junkang Wu, Wentao Shi, Xuezhi Cao, Jiawei Chen, Wenqiang Lei, Fuzheng Zhang, Wei Wu, and Xiangnan He. 2021. Disenkgat: knowledge graph embedding with disentangled graph attention network. In Proceedings of the 30th ACM international conference on information & knowledge management, pages 2140–2149. find
  23. Xiaoyu Kou, Yankai Lin, Yuntao Li, Jiahao Xu, Peng Li, Jie Zhou, and Yan Zhang. 2020. Disene: Disentangling knowledge graph embeddings. arXiv preprint arXiv:2010.14730. — [arXiv:2010.14730](https://arxiv.org/abs/2010.14730) find
  24. A Arun Kumar, Samarth Bhandary, Swathi Gopal Hegde, and Jhinuk Chatterjee. 2025. Knowledge graph applications and multi-relation learning for drug repurposing: A scoping review. Computational Biology and Chemistry, page 108364. find
  25. Jaejun Lee, Chanyoung Chung, and Joyce Jiyoung Whang. 2023. Ingram: Inductive knowledge graph embedding via relation graphs. In International Conference on Machine Learning, pages 18796–18809. PMLR. find
  26. Mingchen Li, Chen Ling, Rui Zhang, and Liang Zhao. 2024. Zero-shot link prediction in knowledge graphs with large language models. In 2024 IEEE International Conference on Data Mining (ICDM), pages 753–760. IEEE. find
  27. Zhao Li, Xin Wang, Jun Zhao, Wenbin Guo, and Jianxin Li. 2025. Hycube: Efficient knowledge hypergraph 3d circular convolutional embedding. IEEE Transactions on Knowledge and Data Engineering. find
  28. Andrew L Maas, Awni Y Hannun, Andrew Y Ng, and 1 others. 2013. Rectifier nonlinearities improve neural network acoustic models. In Proc. icml, volume 30, page 3. Atlanta, GA. find
  29. Kuan Xu, Kuo Yang, Jian Liu, Xiangkui Lu, Jun Wu, and Xuezhong Zhou. 2025. Kgcrr: An effective metric-driven knowledge graph completion framework by designing a novel upper bound function with adaptive approximation to reciprocal rank. In Proceedings of the AAAI Conference on Artificial Intelligence… find
  30. Bishan Yang, Wen-tau Yih, Xiaodong He, Jianfeng Gao, and Li Deng. 2014. Embedding entities and relations for learning and inference in knowledge bases. arXiv preprint arXiv:1412.6575.
  31. Jing Yang, Xiaowen Jiang, Yuan Gao, Laurence T Yang, and Jieming Yang. 2024. Generalize to fully unseen graphs: Learn transferable hyper-relation structures for inductive link prediction. In Proceedings of the 32nd ACM International Conference on Multimedia, pages 1274–1282. find
  32. Hang Yu and Haoyi Yu. 2025. Enhancing zero-shot knowledge graph relation prediction through large language models and contrastive learning. In Companion Proceedings of the ACM on Web Conference 2025, pages 1490–1494. find
  33. Laurens van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-sne. Journal of machine learning research, 9(Nov):2579–2605. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/xu-2026-structure-aware-zero-shot-relational/