Elamur Research⌕
☰
home / papers / 2026 / MAKI: Multi-layer Aligned Knowledge Injection for Structure-aware Knowledge Graph Completion with Large Language Models

MAKI: Multi-layer Aligned Knowledge Injection for Structure-aware Knowledge Graph Completion with Large Language Models

Paper · 2026 · Zhiwen Xie, Xin Wang, Guangyou Zhou et al.

Summary

MAKI addresses a weakness in LLM-based knowledge graph completion (KGC). Current LLM approaches use little of the graph's topology. Methods such as KoPA and SAT add structure through RotatE-style embeddings, but those embeddings live in a different vector space from the LLM and cannot represent entities unseen in training.

MAKI works in three steps:

  • It feeds entity and relation names through the LLM and keeps the last-token hidden states from several of the top layers.
  • It passes these states to a GNN in which each layer is tied to one LLM layer, so the structural features stay in the LLM's semantic space.
  • It injects each GNN layer's output back into the matching LLM layer through a gate, and the LLM then scores candidate triples.

The authors test MAKI on three transductive benchmarks and eight versions of inductive datasets. They report that it beats state-of-the-art baselines. Because entities are represented from their text, the method extends to new entities, which suits evolving knowledge graphs.

Key points

  • Ablations show every component matters. Without LLM-aligned initialization, CoDeX-S accuracy falls from 87.01% to 83.32%, and MRR on FB15k237 v1 falls from 86.45% to 85.39%. Removing the GNN hurts even more, and replacing the gate with plain addition also lowers scores.
  • Results are stable across backbones (LLaMA3.1-8B, LLaMA2-7B, Qwen2-7B, Alpaca-7B). On CoDeX-S, F1 ranges from 86.43% to 87.24% and accuracy from 86.28% to 87.01%; LLaMA3.1-8B scores highest.
  • Six aligned GNN layers (k = 6) work best. Fewer layers inject too little structure, and more layers bring no further gain.
  • The main limitations are cost. Each candidate triple must be scored separately, which slows large-scale link prediction, and fine-tuning the LLM together with the GNN alignment is computationally expensive.

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

🎓University of MacauMacao · China43 works · 2,368 citations🎓Central China Normal UniversityWuhan · China3 works · 1 citation

 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 Zhiwen Xie, Xin Wang, Guangyou Zhou and Derek F. Wong. 2026. MAKI: Multi-layer Aligned Knowledge Injection for Structure-aware Knowledge Graph Completion with Large Language Models. 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.1423

@inproceedings{xie2026maki,
  title = {{MAKI: Multi-layer Aligned Knowledge Injection for Structure-aware Knowledge Graph Completion with Large Language Models}},
  author = {Zhiwen Xie, Xin Wang, Guangyou Zhou and Derek F. Wong},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1423},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1423}
}

References

33 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Olivier Bodenreider. 2004. The unified medical language system (UMLS): integrating biomedical terminology. Nucleic Acids Res., 32(Database-Issue):267– 270. find
  2. Antoine Bordes, Nicolas Usunier, Alberto García-Durán, Jason Weston, and Oksana Yakhnenko. 2013. Translating embeddings for modeling multi-relational data. In Advances in Neural Information Processing Systems 26: 27th Annual Conference on Neural Information Processing Systems 2013., pages 2787–2795.
  3. Linlin Chao, Jianshan He, Taifeng Wang, and Wei Chu. 2021. PairRE: Knowledge graph embeddings via paired relation vectors. In Proceedings of the 59th Chen Chen, Yufei Wang, Aixin Sun, Bing Li, and Kwok-Yan Lam. 2023. Dipping plms sauce: Bridging structure and text for effective knowledge graph compl… find
  4. Mingyang Chen, Wen Zhang, Yushan Zhu, Hongting Zhou, Zonggang Yuan, Changliang Xu, and Huajun Chen. 2022. Meta-knowledge transfer for inductive knowledge graph embedding. In SIGIR ’22: The 45th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 927–937. AC… find
  5. Tim Dettmers, Pasquale Minervini, Pontus Stenetorp, and Sebastian Riedel. 2018. Convolutional 2D knowledge graph embeddings. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, (AAAI-18), the 30th innovative Applications of Artificial Intelligence (IAAI-18), and the 8th A… find
  6. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,…
  7. Mor Geva, Jasmijn Bastings, Katja Filippova, and Amir Globerson. 2023. Dissecting recall of factual associations in auto-regressive language models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, EMNLP 2023, Singapore, December 6-10, 2023, pages 12216–1223… find
  8. Lingbing Guo, Zhongpu Bo, Zhuo Chen, Yichi Zhang, Jiaoyan Chen, Yarong Lan, Mengshu Sun, Zhiqiang Zhang, Yangyifei Luo, Qian Li, Qiang Zhang, Wen Zhang, and Huajun Chen. 2024. Mkgl: Mastery of a three-word language. In Advances in Neural Information Processing Systems, volume 37, pages 140509–140534… find
  9. Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. Lora: Low-rank adaptation of large language models. In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022. OpenReview.net.
  10. Junsik Kim, Jinwook Park, and Kangil Kim. 2025. RSCF: Relation-semantics consistent filter for entity embedding of knowledge graph. In Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 12320–12336, Vienna, Austria. Association for… find
  11. Diederik P. Kingma and Jimmy Ba. 2015. Adam: A method for stochastic optimization. In 3rd International Conference on Learning Representations.
  12. Stanley Kok and Pedro M. Domingos. 2007. Statistical predicate invention. In Machine Learning, Proceedings of the Twenty-Fourth International Conference (ICML 2007), Corvallis, Oregon, USA, June 20-24, 2007, volume 227 of ACM International Conference Proceeding Series, pages 433–440. ACM. find
  13. Yu Liu, Yanan Cao, Xixun Lin, Yanmin Shang, Shi Wang, and Shirui Pan. 2025. Enhancing large language model for knowledge graph completion via structure-aware alignment-tuning. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, Suzhou, China. Association for Co… find
  14. Yu Liu, Yanan Cao, Shi Wang, Qingyue Wang, and Guanqun Bi. 2024. Generative models for complex logical reasoning over knowledge graphs. In Proceedings of the 17th ACM International Conference on Web Search and Data Mining, WSDM ’24, page 492–500, New York, NY, USA. Association for Computing Machiner… find
  15. Xiao Long, Liansheng Zhuang, Aodi Li, Minghong Yao, and Shafei Wang. 2025. EPERM: an evidence path enhanced reasoning model for knowledge graph question and answering. In AAAI-25, Sponsored by the Association for the Advancement of Artificial Intelligence, February 25 - March 4, 2025, Philadelphia,… find
  16. 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. In Findings of the Association for Computational Linguistics: ACL 2022, Dublin, Ireland, May 22-27,… find
  17. Sijie Mai, Shuangjia Zheng, Yuedong Yang, and Haifeng Hu. 2021. Communicative message passing for inductive relation reasoning. In Thirty-Fifth AAAI Conference on Artificial Intelligence, AAAI 2021, Thirty-Third Conference on Innovative Applications of Artificial Intelligence, IAAI 2021, The Elevent… find
  18. Fanshen Meng, Zhenhua Meng, Ru Jin, Rongheng Lin, and Budan Wu. 2025. DOGE: llms-enhanced hyper-knowledge graph recommender for multimodal recommendation. In AAAI-25, Sponsored by the Association for the Advancement of Artificial Intelligence, February 25 - March 4, 2025, Philadelphia, PA, USA, page… find
  19. OpenAI. 2023. GPT-4 technical report. abs/2303.08774.
  20. CoRR, Moritz Plenz and Anette Frank. 2024. Graph language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 4477–4494, Bangkok, Thailand. Association for Computational Linguistics. find
  21. Tara Safavi and Danai Koutra. 2020. Codex: A comprehensive knowledge graph completion benchmark. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020, Online, November 16-20, 2020, pages 8328– 8350. Association for Computational Linguistics. find
  22. Zhiqing Sun, Zhi-Hong Deng, Jian-Yun Nie, and Jian Tang. 2019. RotatE: Knowledge graph embedding by relational rotation in complex space. In 7th International Conference on Learning Representations.
  23. Rohan Taori, Ishaan Gulrajani, Tianyi Zhang, Yann Dubois, Xuechen Li, Carlos Guestrin, Percy Liang, and Tatsunori B. Hashimoto. 2023. Stanford alpaca: An instruction-following llama model. https://github.com/tatsu-lab/stanford_alpaca. ↗
  24. Komal K. Teru, Etienne G. Denis, and William L. Hamilton. 2020. Inductive relation prediction by subgraph reasoning. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event, volume 119 of Proceedings of Machine Learning Research, pages 9448–… find
  25. Kristina Toutanova, Danqi Chen, Patrick Pantel, Hoifung Poon, Pallavi Choudhury, and Michael Gamon. 2015. Representing text for joint embedding of text and knowledge bases. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 1499–1509.
  26. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, Aurélien Rodriguez, Armand Joulin, Edouard Grave, and Guillaume Lample. 2023a. Llama: Open and efficient foundation language models. CoRR, a…
  27. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, Dan Bikel, Lukas Blecher, Cristian Canton-Ferrer, Moya Chen, Guillem Cucurull, David Esiobu, Jude Fernandes, Jeremy Fu, Wenyin Fu, and 49 others… — Llama 2: Open Foundation and Fine-Tuned Chat Models
  28. Théo Trouillon, Johannes Welbl, Sebastian Riedel, Éric Gaussier, and Guillaume Bouchard. 2016. Complex embeddings for simple link prediction. In Proceedings of the 33nd International Conference on Machine Learning, pages 2071–2080.
  29. Liang Wang, Wei Zhao, Zhuoyu Wei, and Jingming Liu. 2022. Simkgc: Simple contrastive knowledge graph completion with pre-trained language models. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), ACL 2022, pages 4281–4294. Association… find
  30. Weihua Wang, Qiuyu Liang, Feilong Bao, and Guanglai Gao. 2025. Distance-adaptive quaternion knowledge graph embedding with bidirectional rotation. In Proceedings of the 31st International Conference on Computational Linguistics, COLING 2025, Abu Dhabi, UAE, January 19-24, 2025, pages 4219–4231. Asso… find
  31. An Yang, Baosong Yang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Zhou, Chengpeng Li, Chengyuan Li, Dayiheng Liu, Fei Huang, Guanting Dong, Haoran Wei, Huan Lin, Jialong Tang, Jialin Wang, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Ma, and 40 others. 2024. Qwen2 technical report. arXiv preprint arXi… find
  32. Bishan Yang, Wen-tau Yih, Xiaodong He, Jianfeng Gao, and Li Deng. 2015. Embedding entities and relations for learning and inference in knowledge bases. In 3rd International Conference on Learning Representations.
  33. Liang Yao, Chengsheng Mao, and Yuan Luo. 2019. KG-BERT: BERT for knowledge graph completion. CoRR, abs/1909.03193. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/xie-2026-maki-multi-layer-aligned-knowledge/