Elamur Research⌕
☰
home / papers / 2026 / AnchorAlign: A Novel Anchor Alignment-enhanced Generative Method for Joint Named Entity Recognition and Relation Extraction

AnchorAlign: A Novel Anchor Alignment-enhanced Generative Method for Joint Named Entity Recognition and Relation Extraction

Paper · 2026 · Xiaolong Weng, Yuanyun Zhou, Boyu Qiu et al.

Summary

Generative models now dominate information extraction, but joint systems that both recognize named entities (NER) and extract the relations among them (RE) often produce outputs that disagree. Entities found by NER may not match the arguments used in relations, and relations that share an entity are generated without regard to one another. Earlier alignment schemes typically link an NER entity only to its closest relation mention, which introduces bias.

AnchorAlign, built on the BARTNER-style generative architecture, addresses this with "anchor entities": entities shared between the NER and RE outputs, or shared across several relations. Using anchors, the authors unify existing structured output paradigms and derive two new ones. The model uses a shared encoder and three decoders: one for NER, one forward subject-grouped RE decoder and one reverse object-grouped RE decoder. Training aligns their output probability distributions on anchor tokens. At inference, anchor-guided constraints prune and merge predictions.

On five benchmarks, the authors report that AnchorAlign beats SOTA baselines. The approach is potentially extensible to joint NER and event extraction.

Key points

  • Two kinds of anchors:
    • entity-relation anchors, which are entities shared by NER and RE;
    • relation-relation anchors, which are shared head or tail entities and shared triplets.
  • Label encoding: labels such as <PER> and <WORK FOR> are single special tokens. Each is initialized by mean-pooling the embeddings of its label text, which removes bracket boundary markers and shortens output sequences.
  • Training losses:
    • a KL-based NER–RE alignment loss, averaged over the forward and reverse RE decoders;
    • an additional forward–reverse RE consistency term, intended to reduce generation-order bias.
  • Datasets and cost: evaluation datasets include ACE05, SciERC and the Chinese medical Text2DT. The three-decoder model has 914M parameters (BART-Large backbone). Its decoders run in parallel, but it is still slower than BiSPN and PL-Marker.
  • Limitations: alignment sharpens entity boundaries but does not fix semantically wrong entities. Relation-type errors and head-tail reversal persist, which the authors attribute to missing domain knowledge.

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

🎓Harbin Institute of TechnologyHarbin · China185 works · 29,140 citations🔬Peng Cheng LaboratoryShenzhen · China47 works · 2,699 citations

 Fields, topics and concepts

📚 Shelves 5The 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 worksComputer Vision and Pattern Recognition · 3,486 works

🧭 Topics 2Research subjects and languages the work is about.Information extraction · 175 worksNamed entity recognition · 214 works

Cite

Cite as Xiaolong Weng, Yuanyun Zhou, Boyu Qiu, Zehua Wang, Ying Xiong and Buzhou Tang. 2026. AnchorAlign: A Novel Anchor Alignment-enhanced Generative Method for Joint Named Entity Recognition and Relation Extraction. 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.1898

@inproceedings{weng2026anchoralign,
  title = {{AnchorAlign: A Novel Anchor Alignment-enhanced Generative Method for Joint Named Entity Recognition and Relation Extraction}},
  author = {Xiaolong Weng, Yuanyun Zhou, Boyu Qiu, Zehua Wang, Ying Xiong and Buzhou Tang},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1898},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1898}
}

References

24 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Masaki Asada and Makoto Miwa. 2025. Improving relation extraction by sequence-to-sequence-based dependency parsing pre-training. In Proceedings of the 31st International Conference on Computational Linguistics, pages 7099–7105, Abu Dhabi, UAE. Association for Computational Linguistics. find
  2. Pere-Lluís Huguet Cabot and Roberto Navigli. 2021. Rebel: Relation extraction by end-to-end language generation. In Findings of the association for computational linguistics: emnlp 2021, pages 2370–2381. find
  3. Harsha Gurulingappa, Abdul Mateen Rajput, Angus Roberts, Juliane Fluck, Martin Hofmann-Apitius, and Luca Toldo. 2012. Development of a benchmark corpus to support the automatic extraction of drug-related adverse effects from medical case reports. find
  4. Journal of Biomedical Informatics, 45(5):885–892. Text Mining and Natural Language Processing in Pharmacogenomics. find
  5. Yuxin He and Buzhou Tang. 2023. Bispn: Generating entity set and relation set coherently in one pass. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 2066–2077. find
  6. Martin Josifoski, Nicola De Cao, Maxime Peyrard, Fabio Petroni, and Robert West. 2022. GenIE: Generative information extraction. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4626–4643, Seattle… find
  7. Hyeonseok Kang, Hyein Seo, Jeesu Jung, Sangkeun Jung, Du-Seong Chang, and Riwoo Chung. 2024. Guidance-based prompt data augmentation in specialized domains for named entity recognition. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers… find
  8. Peng Li, Tianxiang Sun, Qiong Tang, Hang Yan, Yuanbin Wu, Xuanjing Huang, and Xipeng Qiu. 2023. CodeIE: Large code generation models are better few-shot information extractors. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1…
  9. Zixuan Li, Yutao Zeng, Yuxin Zuo, Weicheng Ren, Wenxuan Liu, Miao Su, Yucan Guo, Yantao Liu, Xiang Li, Zhilei Hu, and 1 others. 2024. Knowcoder: Coding structured knowledge into llms for universal information extraction. arXiv preprint arXiv:2403.07969.
  10. Jie Lou, Yaojie Lu, Dai Dai, Wei Jia, Hongyu Lin, Xianpei Han, Le Sun, and Hua Wu. 2023. Universal information extraction as unified semantic matching. In Proceedings of the AAAI conference on Artificial Intelligence, volume 37, pages 13318–13326. find
  11. Yaojie Lu, Qing Liu, Dai Dai, Xinyan Xiao, Hongyu Lin, Xianpei Han, Le Sun, and Hua Wu. 2022. Unified structure generation for universal information extraction. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5755–5772, Dublin… find
  12. Yi Luan, Luheng He, Mari Ostendorf, and Hannaneh Hajishirzi. 2018. Multi-task identification of entities, relations, and coreference for scientific knowledge graph construction. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing.
  13. Giovanni Paolini, Ben Athiwaratkun, Jason Krone, Jie Ma, Alessandro Achille, Rishita Anubhai, Cicero Nogueira dos Santos, Bing Xiang, and Stefano Soatto. 2021. Structured prediction as translation between augmented natural languages. arXiv preprint arXiv:2101.05779. find
  14. Sebastian Riedel, Limin Yao, and Andrew McCallum. 2010. Modeling Relations and Their Mentions without Labeled Text, page 148–163. find
  15. Oscar Sainz, Iker García-Ferrero, Rodrigo Agerri, Oier Lopez de Lacalle, German Rigau, and Eneko Agirre. 2023. Gollie: Annotation guidelines improve zero-shot information-extraction. arXiv preprint arXiv:2310.03668.
  16. C. Walker and Linguistic Data Consortium. 2005. find
  17. ACE 2005 Multilingual Training Corpus. LDC corpora. Linguistic Data Consortium. find
  18. Chenguang Wang, Xiao Liu, Zui Chen, Haoyun Hong, Jie Tang, and Dawn Song. 2022. DeepStruct: Pre-training of language models for structure prediction. In Findings of the Association for Computational Linguistics: ACL 2022, pages 803–823, Dublin, Ireland. Association for Computational Linguistics. find
  19. Xiao Wang, Weikang Zhou, Can Zu, Han Xia, Tianze Chen, Yuansen Zhang, Rui Zheng, Junjie Ye, Qi Zhang, Tao Gui, and 1 others. 2023. Instructuie: Multi-task instruction tuning for unified information extraction. arXiv preprint arXiv:2304.08085.
  20. Xinglin Xiao, Yijie Wang, Nan Xu, Yuqi Wang, Hanxuan Yang, Minzheng Wang, Yin Luo, Lei Wang, Wenji Mao, and Daniel Zeng. 2023. Yayi-uie: A chat-enhanced instruction tuning framework for universal information extraction. arXiv preprint arXiv:2312.15548. — [arXiv:2312.15548](https://arxiv.org/abs/2312.15548) find
  21. Hang Yan, Tao Gui, Junqi Dai, Qipeng Guo, Zheng Zhang, and Xipeng Qiu. 2021. A unified generative framework for various NER subtasks. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing…
  22. Deming Ye, Yankai Lin, Peng Li, and Maosong Sun. 2022. Packed levitated marker for entity and relation extraction. In Proceedings of the 60th annual meeting of the association for computational linguistics (volume 1: long papers), pages 4904–4917.
  23. Hengyi Zheng, Rui Wen, Xi Chen, Yifan Yang, Yunyan Zhang, Ziheng Zhang, Ningyu Zhang, Bin Qin, Xu Ming, and Yefeng Zheng. 2021. PRGC: Potential relation and global correspondence based joint relational triple extraction. In Proceedings of the 59th Annual Meeting of the Association for Computational… find
  24. Wei Zhu, Wenfeng Li, Xiaoling Wang, Wendi Ji, Yuanbin Wu, Jin Chen, Liang Chen, and Buzhou Tang. 2023. Extracting Decision Trees from Medical Texts: An Overview of the Text2DT Track in CHIP2022, page 89–102. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/weng-2026-anchoralign-novel-anchor-alignment-enhanced/