---
title: "MuSe: Multi-Stage Graph Reasoning via Vision-Language Models"
description: "The authors also release LGVLQA, a multimodal benchmark with bigger and more complex graphs than prior datasets such as… 51 references."
authors: ["Guanyu Wang", "Xu Chu", "Zhijie Tan", "Xinrong Chen", "Tong Mo", "Weiping Li"]
year: 2026
venue: "Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)"
source: "https://aclanthology.org/2026.acl-long.476/"
doi: "10.18653/v1/2026.acl-long.476"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/wang-2026-muse-multi-stage-graph-reasoning/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/wang-2026-muse-multi-stage-graph-reasoning/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/37/) / MuSe: Multi-Stage Graph Reasoning via Vision-Language Models

# MuSe: Multi-Stage Graph Reasoning via Vision-Language Models

**Authors:** [Guanyu Wang](https://md.research.elamur.ai/free/en/authors/guanyu-wang/), [Xu Chu](https://md.research.elamur.ai/free/en/authors/xu-chu/), [Zhijie Tan](https://md.research.elamur.ai/free/en/authors/zhijie-tan/), Xinrong Chen, [Tong Mo](https://md.research.elamur.ai/free/en/authors/tong-mo/) and [Weiping Li](https://md.research.elamur.ai/free/en/authors/weiping-li/) · **Year:** 2026 · **Venue:** [Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)](https://md.research.elamur.ai/free/en/venues/acl/) ([ACL 2026](https://md.research.elamur.ai/free/en/venues/acl/2026/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.acl-long.476/) · [PDF](https://aclanthology.org/2026.acl-long.476.pdf) · [DOI 10.18653/v1/2026.acl-long.476](https://doi.org/10.18653/v1/2026.acl-long.476) · [OpenAlex](https://openalex.org/W7166813287)

**Citations:** 0 citations (OpenAlex, 2026-10-14)

## Summary
Graph reasoning is usually handled by GNNs or graph transformers, which need task-specific training. LLMs that read graphs serialized as text struggle to keep two-dimensional structure and multi-hop relations. Vision-language methods such as GITA draw the entire graph as one picture, but large graphs then become cluttered, with crossing or overlapping edges and irrelevant regions.

MuSe treats graph reasoning as a multi-turn dialog. In each round, the VLM chooses a node and hop radius to sample. A visualizer returns an image of that local subgraph, and the model collects evidence until it can answer.

Training has two phases:
- **SFT** on GPT-generated dialogs teaches sampling and local reasoning. The dialogs were produced with gold answers supplied.
- **GRPO** then sharpens the sampling policy and keeps conversations short.

The authors also release LGVLQA, a multimodal benchmark with bigger and more complex graphs than prior datasets such as GVLQA. MuSe beats leading LLM and VLM baselines across graph reasoning tasks. This suggests that focused, incremental visualization scales better than rendering the full graph.

## Key points
- The model acts through tagged steps: `<think>`, `<sample>` with calls like `sample(G, node, hops)`, and `<answer>`. Each sampled subgraph image is added to the dialog history.
- Training-split ablation on GVLQA (average accuracy):
  - 80% SFT followed by 20% GRPO: 60.7
  - SFT on all data: 59.1
  - 20% cold-start SFT followed by 80% GRPO: 34.8
- With the 80/20 split, task-level accuracy is MaxFlow 22.6, SP 35.3 and TS 42.7.
- Varying the maximum interaction rounds K gives average accuracy of 48.4 (K=2), 56.8 (K=3), 60.0 (K=4), 60.7 (K=5) and 60.9 (K=10). About five rounds are enough, and more rounds mostly add training and inference cost.
- Stated limitations:
  - The GPT-written reasoning traces can oversimplify complex problems.
  - The method is currently limited to plain graph reasoning; extensions to text-attributed graphs and knowledge graphs are planned.

*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

- [Peking University](https://md.research.elamur.ai/free/en/universities/peking-university/) — University, Beijing, China · 244 works · 39,359 citations

📚 Shelves (3): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 2,869 works

🧭 Topics (3): [🧠 Language models & pretraining](https://md.research.elamur.ai/free/en/topics/language-models/) · 3,320 works · [🧮 Reasoning](https://md.research.elamur.ai/free/en/topics/reasoning/) · 1,426 works · [🖼️ Vision & language](https://md.research.elamur.ai/free/en/topics/multimodal/) · 657 works

## Cite

Cite as: Guanyu Wang, Xu Chu, Zhijie Tan, Xinrong Chen, Tong Mo and Weiping Li. 2026. MuSe: Multi-Stage Graph Reasoning via Vision-Language Models. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{wang2026muse,
  title = {{MuSe: Multi-Stage Graph Reasoning via Vision-Language Models}},
  author = {Guanyu Wang, Xu Chu, Zhijie Tan, Xinrong Chen, Tong Mo and Weiping Li},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.476},
  url = {https://doi.org/10.18653/v1/2026.acl-long.476}
}
```

- [← previous MURAD: A Large-Scale Multi-Domain Unified Reverse Arabic Dictionary Dataset](https://md.research.elamur.ai/free/en/papers/sibaee-2026-murad-large-scale-multi-domain/)
- [next → MUSIA: Multilingual Story Illustration Corpus for Cross-Cultural Alignment and Generation](https://md.research.elamur.ai/free/en/papers/tewari-2026-musia-multilingual-story-illustration-corpus/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/37/)

## References

51 references, 16 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhaohai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, and 8 others. 2025. [Qwen2.5-vl technical report](https://md.research.elamur.ai/free/en/papers/bai-2025-qwen2-5-vl-technical/). Preprint, arXiv:2502.13…
2. Peter W Battaglia, Jessica B Hamrick, Victor Bapst, Alvaro Sanchez-Gonzalez, Vinicius Zambaldi, Mateusz Malinowski, Andrea Tacchetti, David Raposo, Adam Santoro, Ryan Faulkner, and 1 others. 2018. [Relational inductive biases, deep learning, and graph networks](https://md.research.elamur.ai/free/en/papers/battaglia-2018-relational-inductive-biases-deep-learning/). arXiv preprint arXiv:1806.01261.
3. Antoine Bordes, Nicolas Usunier, Alberto Garcia-Duran, Jason Weston, and Oksana Yakhnenko. 2013. [Translating embeddings for modeling multi-relational data](https://md.research.elamur.ai/free/en/papers/bordes-2013-translating-embeddings-modeling-multi-relational/). Advances in neural information processing systems, 26.
4. Qi Cao, Huawei Shen, Jinhua Gao, Bingzheng Wei, and Xueqi Cheng. 2020. Popularity prediction on social platforms with coupled graph neural networks. In Proceedings of the 13th international conference on web search and data mining, pages 70–78. [find](https://scholar.google.com/scholar?q=Qi+Cao%2C+Huawei+Shen%2C+Jinhua+Gao%2C+Bingzheng+Wei%2C+and+Xueqi+Cheng%2E+2020%2E+Popularity+prediction+on+social+platforms+with+coupled+graph+neural+networks%2E+In+Proceedings+of+the+13th+international+conference)
5. Dexiong Chen, Leslie O’Bray, and Karsten Borgwardt. 2022. Structure-aware transformer for graph representation learning. In International conference on machine learning, pages 3469–3489. PMLR. [find](https://scholar.google.com/scholar?q=Dexiong+Chen%2C+Leslie+O%E2%80%99Bray%2C+and+Karsten+Borgwardt%2E+2022%2E+Structure%2Daware+transformer+for+graph+representation+learning%2E+In+International+conference+on+machine+learning%2C+pages+3469%E2%80%933489%2E+PMLR%2E)
6. Nuo Chen, Yuhan Li, Jianheng Tang, and Jia Li. 2024a. Graphwiz: An instruction-following language model for graph computational problems. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 353–364. [find](https://scholar.google.com/scholar?q=Nuo+Chen%2C+Yuhan+Li%2C+Jianheng+Tang%2C+and+Jia+Li%2E+2024a%2E+Graphwiz%3A+An+instruction%2Dfollowing+language+model+for+graph+computational+problems%2E+In+Proceedings+of+the+30th+ACM+SIGKDD+Conference+on+Knowledge+)
7. Runjin Chen, Tong Zhao, Ajay Jaiswal, Neil Shah, and Zhangyang Wang. 2024b. Llaga: Large language and graph assistant. arXiv preprint arXiv:2402.08170. — [arXiv:2402.08170](https://arxiv.org/abs/2402.08170) [find](https://scholar.google.com/scholar?q=Runjin+Chen%2C+Tong+Zhao%2C+Ajay+Jaiswal%2C+Neil+Shah%2C+and+Zhangyang+Wang%2E+2024b%2E+Llaga%3A+Large+language+and+graph+assistant%2E+arXiv+preprint+arXiv%3A2402%2E08170%2E+%E2%80%94+%5BarXiv%3A2402%2E08170%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2402%2E)
8. Zhe Chen, Weiyun Wang, Yue Cao, Yangzhou Liu, Zhangwei Gao, Erfei Cui, Jinguo Zhu, Shenglong Ye, Hao Tian, Zhaoyang Liu, Lixin Gu, Xuehui Wang, Qingyun Li, Yiming Ren, Zixuan Chen, Jiapeng Luo, Jiahao Wang, Tan Jiang, Bo Wang, and 23 others. 2025. [Expanding performance boundaries of open-source mult…](https://md.research.elamur.ai/free/en/papers/chen-2024-expanding-performance-boundaries-open/)
9. Xu Chu, Hanlin Xue, Zhijie Tan, Bingce Wang, Tong Mo, and Weiping Li. 2025. Graphsos: Graph sampling and order selection to help llms understand graphs better. arXiv preprint arXiv:2501.14427. — [arXiv:2501.14427](https://arxiv.org/abs/2501.14427) [find](https://scholar.google.com/scholar?q=Xu+Chu%2C+Hanlin+Xue%2C+Zhijie+Tan%2C+Bingce+Wang%2C+Tong+Mo%2C+and+Weiping+Li%2E+2025%2E+Graphsos%3A+Graph+sampling+and+order+selection+to+help+llms+understand+graphs+better%2E+arXiv+preprint+arXiv%3A2501%2E14427%2E+%E2%80%94+%5BarXi)
10. Bahare Fatemi, Jonathan Halcrow, and Bryan Perozzi. 2023. Talk like a graph: Encoding graphs for large language models. arxiv 2023. arXiv preprint arXiv:2310.04560. [find](https://scholar.google.com/scholar?q=Bahare+Fatemi%2C+Jonathan+Halcrow%2C+and+Bryan+Perozzi%2E+2023%2E+Talk+like+a+graph%3A+Encoding+graphs+for+large+language+models%2E+arxiv+2023%2E+arXiv+preprint+arXiv%3A2310%2E04560%2E)
11. Emden R Gansner and Stephen C North. 2000. An open graph visualization system and its applications to software engineering. Software: practice and experience, 30(11):1203–1233. [find](https://scholar.google.com/scholar?q=Emden+R+Gansner+and+Stephen+C+North%2E+2000%2E+An+open+graph+visualization+system+and+its+applications+to+software+engineering%2E+Software%3A+practice+and+experience%2C+30%2811%29%3A1203%E2%80%931233%2E)
12. Justin Gilmer, Samuel S Schoenholz, Patrick F Riley, Oriol Vinyals, and George E Dahl. 2017. [Neural message passing for quantum chemistry](https://md.research.elamur.ai/free/en/papers/gilmer-2017-neural-message-passing-quantum-chemistry/). In International conference on machine learning, pages 1263–1272. Pmlr.
13. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, and 1 others. 2024. [The llama 3 herd of models](https://md.research.elamur.ai/free/en/papers/grattafiori-2024-llama-3-herd-models/). arXiv preprint arXiv:2407.21783.
14. Jiayan Guo, Lun Du, Hengyu Liu, Mengyu Zhou, Xinyi He, and Shi Han. 2023. Gpt4graph: Can large language models understand graph structured data? an empirical evaluation and benchmarking. arXiv preprint arXiv:2305.15066. [find](https://scholar.google.com/scholar?q=Jiayan+Guo%2C+Lun+Du%2C+Hengyu+Liu%2C+Mengyu+Zhou%2C+Xinyi+He%2C+and+Shi+Han%2E+2023%2E+Gpt4graph%3A+Can+large+language+models+understand+graph+structured+data%3F+an+empirical+evaluation+and+benchmarking%2E+arXiv+preprin)
15. Will Hamilton, Zhitao Ying, and Jure Leskovec. 2017. [Inductive representation learning on large graphs](https://md.research.elamur.ai/free/en/papers/hamilton-2017-inductive-representation-learning-large-graphs/). Advances in neural information processing systems, 30.
16. Xiangnan He, Kuan Deng, Xiang Wang, Yan Li, Yongdong Zhang, and Meng Wang. 2020. Lightgcn: Simplifying and powering graph convolution network for recommendation. In Proceedings of the 43rd International ACM SIGIR conference on research and development in Information Retrieval, pages 639–648. [find](https://scholar.google.com/scholar?q=Xiangnan+He%2C+Kuan+Deng%2C+Xiang+Wang%2C+Yan+Li%2C+Yongdong+Zhang%2C+and+Meng+Wang%2E+2020%2E+Lightgcn%3A+Simplifying+and+powering+graph+convolution+network+for+recommendation%2E+In+Proceedings+of+the+43rd+Internation)
17. Xiangnan He, Lizi Liao, Hanwang Zhang, Liqiang Nie, Xia Hu, and Tat-Seng Chua. 2017. Neural collaborative filtering. In Proceedings of the 26th international conference on world wide web, pages 173–182. [find](https://scholar.google.com/scholar?q=Xiangnan+He%2C+Lizi+Liao%2C+Hanwang+Zhang%2C+Liqiang+Nie%2C+Xia+Hu%2C+and+Tat%2DSeng+Chua%2E+2017%2E+Neural+collaborative+filtering%2E+In+Proceedings+of+the+26th+international+conference+on+world+wide+web%2C+pages+173%E2%80%9318)
18. Xiaoxin He, Xavier Bresson, Thomas Laurent, Adam Perold, Yann LeCun, and Bryan Hooi. 2024. Harnessing explanations: Llm-to-lm interpreter for enhanced text-attributed graph representation learning. In International Conference on Learning Representations, volume 2024, pages 5711–5732. [find](https://scholar.google.com/scholar?q=Xiaoxin+He%2C+Xavier+Bresson%2C+Thomas+Laurent%2C+Adam+Perold%2C+Yann+LeCun%2C+and+Bryan+Hooi%2E+2024%2E+Harnessing+explanations%3A+Llm%2Dto%2Dlm+interpreter+for+enhanced+text%2Dattributed+graph+representation+learning%2E+In)
19. Chao Huang, Huance Xu, Yong Xu, Peng Dai, Lianghao Xia, Mengyin Lu, Liefeng Bo, Hao Xing, Xiaoping Lai, and Yanfang Ye. 2021. Knowledge-aware coupled graph neural network for social recommendation. In Proceedings of the AAAI conference on artificial intelligence, volume 35, pages 4115–4122. [find](https://scholar.google.com/scholar?q=Chao+Huang%2C+Huance+Xu%2C+Yong+Xu%2C+Peng+Dai%2C+Lianghao+Xia%2C+Mengyin+Lu%2C+Liefeng+Bo%2C+Hao+Xing%2C+Xiaoping+Lai%2C+and+Yanfang+Ye%2E+2021%2E+Knowledge%2Daware+coupled+graph+neural+network+for+social+recommendation%2E+In)
20. Bo Jiang, Ziyan Zhang, Doudou Lin, Jin Tang, and Bin Luo. 2019. Semi-supervised learning with graph learning-convolutional networks. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 11313–11320. [find](https://scholar.google.com/scholar?q=Bo+Jiang%2C+Ziyan+Zhang%2C+Doudou+Lin%2C+Jin+Tang%2C+and+Bin+Luo%2E+2019%2E+Semi%2Dsupervised+learning+with+graph+learning%2Dconvolutional+networks%2E+In+Proceedings+of+the+IEEE%2FCVF+conference+on+computer+vision+and+pa)
21. Weisen Jiang, Han Shi, Longhui Yu, Zhengying Liu, Yu Zhang, Zhenguo Li, and James T Kwok. 2023a. Forward-backward reasoning in large language models for mathematical verification. arXiv preprint arXiv:2308.07758. [find](https://scholar.google.com/scholar?q=Weisen+Jiang%2C+Han+Shi%2C+Longhui+Yu%2C+Zhengying+Liu%2C+Yu+Zhang%2C+Zhenguo+Li%2C+and+James+T+Kwok%2E+2023a%2E+Forward%2Dbackward+reasoning+in+large+language+models+for+mathematical+verification%2E+arXiv+preprint+arXiv)
22. Weisen Jiang, Yu Zhang, and James Kwok. 2023b. Effective structured prompting by meta-learning and representative verbalizer. In International Conference on Machine Learning, pages 15186–15199. PMLR. [find](https://scholar.google.com/scholar?q=Weisen+Jiang%2C+Yu+Zhang%2C+and+James+Kwok%2E+2023b%2E+Effective+structured+prompting+by+meta%2Dlearning+and+representative+verbalizer%2E+In+International+Conference+on+Machine+Learning%2C+pages+15186%E2%80%9315199%2E+PMLR%2E)
23. Yehuda Koren, Robert Bell, and Chris Volinsky. 2009. Matrix factorization techniques for recommender systems. Computer, 42(8):30–37. [find](https://scholar.google.com/scholar?q=Yehuda+Koren%2C+Robert+Bell%2C+and+Chris+Volinsky%2E+2009%2E+Matrix+factorization+techniques+for+recommender+systems%2E+Computer%2C+42%288%29%3A30%E2%80%9337%2E)
24. Devin Kreuzer, Dominique Beaini, Will Hamilton, Vincent Létourneau, and Prudencio Tossou. 2021. Rethinking graph transformers with spectral attention. Advances in Neural Information Processing Systems, 34:21618–21629. [find](https://scholar.google.com/scholar?q=Devin+Kreuzer%2C+Dominique+Beaini%2C+Will+Hamilton%2C+Vincent+L%C3%A9tourneau%2C+and+Prudencio+Tossou%2E+2021%2E+Rethinking+graph+transformers+with+spectral+attention%2E+Advances+in+Neural+Information+Processing+Systems)
25. Jure Leskovec, Lars Backstrom, Ravi Kumar, and Andrew Tomkins. 2008. Microscopic evolution of social networks. In Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 462–470. [find](https://scholar.google.com/scholar?q=Jure+Leskovec%2C+Lars+Backstrom%2C+Ravi+Kumar%2C+and+Andrew+Tomkins%2E+2008%2E+Microscopic+evolution+of+social+networks%2E+In+Proceedings+of+the+14th+ACM+SIGKDD+international+conference+on+Knowledge+discovery+and)
26. Hanqing Li, Kiran Sheena Jyothi, Henry Liang, Sharika Mahadevan, and Diego Klabjan. 2025. Zero-shot graph reasoning via retrieval augmented framework with llms. arXiv preprint arXiv:2509.12743. [find](https://scholar.google.com/scholar?q=Hanqing+Li%2C+Kiran+Sheena+Jyothi%2C+Henry+Liang%2C+Sharika+Mahadevan%2C+and+Diego+Klabjan%2E+2025%2E+Zero%2Dshot+graph+reasoning+via+retrieval+augmented+framework+with+llms%2E+arXiv+preprint+arXiv%3A2509%2E12743%2E)
27. Hao Liu, Jiarui Feng, Lecheng Kong, Ningyue Liang, Dacheng Tao, Yixin Chen, and Muhan Zhang. 2023. One for all: Towards training one graph model for all classification tasks. arXiv preprint arXiv:2310.00149. — [arXiv:2310.00149](https://arxiv.org/abs/2310.00149) [find](https://scholar.google.com/scholar?q=Hao+Liu%2C+Jiarui+Feng%2C+Lecheng+Kong%2C+Ningyue+Liang%2C+Dacheng+Tao%2C+Yixin+Chen%2C+and+Muhan+Zhang%2E+2023%2E+One+for+all%3A+Towards+training+one+graph+model+for+all+classification+tasks%2E+arXiv+preprint+arXiv%3A2310)
28. Xiao Liu, Shiyu Zhao, Kai Su, Yukuo Cen, Jiezhong Qiu, Mengdi Zhang, Wei Wu, Yuxiao Dong, and Jie Tang. 2022. Mask and reason: Pre-training knowledge graph transformers for complex logical queries. In Proceedings of the 28th ACM SIGKDD conference on knowledge discovery and data mining, pages 1120–11… [find](https://scholar.google.com/scholar?q=Xiao+Liu%2C+Shiyu+Zhao%2C+Kai+Su%2C+Yukuo+Cen%2C+Jiezhong+Qiu%2C+Mengdi+Zhang%2C+Wei+Wu%2C+Yuxiao+Dong%2C+and+Jie+Tang%2E+2022%2E+Mask+and+reason%3A+Pre%2Dtraining+knowledge+graph+transformers+for+complex+logical+queries%2E+In)
29. Mark EJ Newman. 2003. The structure and function of complex networks. SIAM review, 45(2):167–256. [find](https://scholar.google.com/scholar?q=Mark+EJ+Newman%2E+2003%2E+The+structure+and+function+of+complex+networks%2E+SIAM+review%2C+45%282%29%3A167%E2%80%93256%2E)
30. Bryan Perozzi, Bahare Fatemi, Dustin Zelle, Anton Tsitsulin, Mehran Kazemi, Rami Al-Rfou, and Jonathan Halcrow. 2024. Let your graph do the talking: Encoding structured data for llms. arXiv preprint arXiv:2402.05862. — [arXiv:2402.05862](https://arxiv.org/abs/2402.05862) [find](https://scholar.google.com/scholar?q=Bryan+Perozzi%2C+Bahare+Fatemi%2C+Dustin+Zelle%2C+Anton+Tsitsulin%2C+Mehran+Kazemi%2C+Rami+Al%2DRfou%2C+and+Jonathan+Halcrow%2E+2024%2E+Let+your+graph+do+the+talking%3A+Encoding+structured+data+for+llms%2E+arXiv+preprint+a)
31. Qwen, :, An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, and 25 others. 2025. Qwen2.5 technical report. Preprint, arXiv:2412.15115. [find](https://scholar.google.com/scholar?q=Qwen%2C+%3A%2C+An+Yang%2C+Baosong+Yang%2C+Beichen+Zhang%2C+Binyuan+Hui%2C+Bo+Zheng%2C+Bowen+Yu%2C+Chengyuan+Li%2C+Dayiheng+Liu%2C+Fei+Huang%2C+Haoran+Wei%2C+Huan+Lin%2C+Jian+Yang%2C+Jianhong+Tu%2C+Jianwei+Zhang%2C+Jianxin+Yang%2C+Jiaxi+)
32. Zhihong Shao, Peiyi Wang, Qihao Zhu, Runxin Xu, Junxiao Song, Xiao Bi, Haowei Zhang, Mingchuan Zhang, YK Li, Yang Wu, and 1 others. 2024. [Deepseekmath: Pushing the limits of mathematical reasoning in open language models](https://md.research.elamur.ai/free/en/papers/shao-2024-deepseekmath-pushing-limits-mathematical-reasoning/). arXiv preprint arXiv:2402.03300.
33. Jiabin Tang, Yuhao Yang, Wei Wei, Lei Shi, Lixin Su, Suqi Cheng, Dawei Yin, and Chao Huang. 2024. Graphgpt: Graph instruction tuning for large language models. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 491–500. [find](https://scholar.google.com/scholar?q=Jiabin+Tang%2C+Yuhao+Yang%2C+Wei+Wei%2C+Lei+Shi%2C+Lixin+Su%2C+Suqi+Cheng%2C+Dawei+Yin%2C+and+Chao+Huang%2E+2024%2E+Graphgpt%3A+Graph+instruction+tuning+for+large+language+models%2E+In+Proceedings+of+the+47th+International)
34. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, and 1 others. 2023. [Llama 2: Open foundation and fine-tuned chat models](https://md.research.elamur.ai/free/en/papers/touvron-2023-llama-2-open-foundation-fine/). arXiv preprint arXiv:2307.09288.
35. Petar Veličković, Guillem Cucurull, Arantxa Casanova, Adriana Romero, Pietro Lio, and Yoshua Bengio. 2017. [Graph attention networks](https://md.research.elamur.ai/free/en/papers/veli-vckovic-2017-graph-attention-networks/). arXiv preprint arXiv:1710.10903.
36. Heng Wang, Shangbin Feng, Tianxing He, Zhaoxuan Tan, Xiaochuang Han, and Yulia Tsvetkov. 2023. Can language models solve graph problems in natural language? Advances in Neural Information Processing Systems, 36:30840–30861. [find](https://scholar.google.com/scholar?q=Heng+Wang%2C+Shangbin+Feng%2C+Tianxing+He%2C+Zhaoxuan+Tan%2C+Xiaochuang+Han%2C+and+Yulia+Tsvetkov%2E+2023%2E+Can+language+models+solve+graph+problems+in+natural+language%3F+Advances+in+Neural+Information+Processing+S)
37. Xiang Wang, Tinglin Huang, Dingxian Wang, Yancheng Yuan, Zhenguang Liu, Xiangnan He, and Tat-Seng Chua. 2021. Learning intents behind interactions with knowledge graph for recommendation. In Proceedings of the web conference 2021, pages 878–887. [find](https://scholar.google.com/scholar?q=Xiang+Wang%2C+Tinglin+Huang%2C+Dingxian+Wang%2C+Yancheng+Yuan%2C+Zhenguang+Liu%2C+Xiangnan+He%2C+and+Tat%2DSeng+Chua%2E+2021%2E+Learning+intents+behind+interactions+with+knowledge+graph+for+recommendation%2E+In+Proceedin)
38. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, and 1 others. 2022. [Chain-of-thought prompting elicits reasoning in large language models](https://md.research.elamur.ai/free/en/papers/wei-2022-chain-thought-prompting-elicits-reasoning/). Advances in neural information processing systems, 35:24824– 24837.
39. Yanbin Wei, Shuai Fu, Weisen Jiang, Zejian Zhang, Zhixiong Zeng, Qi Wu, James T. Kwok, and Yu Zhang. 2024a. GITA: graph to visual and textual integration for vision-language graph reasoning. In Advances in Neural Information Processing Systems 38: Annual Conference on Neural Information Processing S… [find](https://scholar.google.com/scholar?q=Yanbin+Wei%2C+Shuai+Fu%2C+Weisen+Jiang%2C+Zejian+Zhang%2C+Zhixiong+Zeng%2C+Qi+Wu%2C+James+T%2E+Kwok%2C+and+Yu+Zhang%2E+2024a%2E+GITA%3A+graph+to+visual+and+textual+integration+for+vision%2Dlanguage+graph+reasoning%2E+In+Advanc)
40. Yanbin Wei, Qiushi Huang, James T Kwok, and Yu Zhang. 2024b. Kicgpt: Large language model with knowledge in context for knowledge graph completion. arXiv preprint arXiv:2402.02389. [find](https://scholar.google.com/scholar?q=Yanbin+Wei%2C+Qiushi+Huang%2C+James+T+Kwok%2C+and+Yu+Zhang%2E+2024b%2E+Kicgpt%3A+Large+language+model+with+knowledge+in+context+for+knowledge+graph+completion%2E+arXiv+preprint+arXiv%3A2402%2E02389%2E)
41. Zonghan Wu, Shirui Pan, Fengwen Chen, Guodong Long, Chengqi Zhang, and Philip S Yu. 2020. [A comprehensive survey on graph neural networks](https://md.research.elamur.ai/free/en/papers/wu-2019-comprehensive-survey-graph-neural-networks/). IEEE transactions on neural networks and learning systems, 32(1):4–24.
42. Keyulu Xu, Weihua Hu, Jure Leskovec, and Stefanie Jegelka. 2018. [How powerful are graph neural networks?](https://md.research.elamur.ai/free/en/papers/xu-2018-powerful-graph-neural-networks/) arXiv preprint arXiv:1810.00826.
43. Junhan Yang, Zheng Liu, Shitao Xiao, Chaozhuo Li, Defu Lian, Sanjay Agrawal, Amit Singh, Guangzhong Sun, and Xing Xie. 2021. Graphform-ers: Gnn-nested transformers for representation learning on textual graph. Advances in Neural Information Processing Systems, 34:28798–28810. [find](https://scholar.google.com/scholar?q=Junhan+Yang%2C+Zheng+Liu%2C+Shitao+Xiao%2C+Chaozhuo+Li%2C+Defu+Lian%2C+Sanjay+Agrawal%2C+Amit+Singh%2C+Guangzhong+Sun%2C+and+Xing+Xie%2E+2021%2E+Graphform%2Ders%3A+Gnn%2Dnested+transformers+for+representation+learning+on+textu)
44. Ruosong Ye, Caiqi Zhang, Runhui Wang, Shuyuan Xu, and Yongfeng Zhang. 2023. Language is all a graph needs. arXiv preprint arXiv:2308.07134. [find](https://scholar.google.com/scholar?q=Ruosong+Ye%2C+Caiqi+Zhang%2C+Runhui+Wang%2C+Shuyuan+Xu%2C+and+Yongfeng+Zhang%2E+2023%2E+Language+is+all+a+graph+needs%2E+arXiv+preprint+arXiv%3A2308%2E07134%2E)
45. Longhui Yu, Weisen Jiang, Han Shi, Jincheng Yu, Zhengying Liu, Yu Zhang, James T Kwok, Zhenguo Li, Adrian Weller, and Weiyang Liu. 2023. [Metamath: Bootstrap your own mathematical questions for large language models](https://md.research.elamur.ai/free/en/papers/yu-2023-metamath-bootstrap-your-own-mathematical/). arXiv preprint arXiv:2309.12284.
46. Jiawei Zhang, Haopeng Zhang, Congying Xia, and Li Sun. 2020. Graph-bert: Only attention is needed for learning graph representations. arxiv 2020. arXiv preprint arXiv:2001.05140. — [arXiv:2001.05140](https://arxiv.org/abs/2001.05140) [find](https://scholar.google.com/scholar?q=Jiawei+Zhang%2C+Haopeng+Zhang%2C+Congying+Xia%2C+and+Li+Sun%2E+2020%2E+Graph%2Dbert%3A+Only+attention+is+needed+for+learning+graph+representations%2E+arxiv+2020%2E+arXiv+preprint+arXiv%3A2001%2E05140%2E+%E2%80%94+%5BarXiv%3A2001%2E05140%5D%28)
47. Yizhuo Zhang, Heng Wang, Shangbin Feng, Zhaoxuan Tan, Xiaochuang Han, Tianxing He, and Yulia Tsvetkov. 2024. Can LLM graph reasoning generalize beyond pattern memorization? In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 2289–2305, Miami, Florida, USA. Association for… [find](https://scholar.google.com/scholar?q=Yizhuo+Zhang%2C+Heng+Wang%2C+Shangbin+Feng%2C+Zhaoxuan+Tan%2C+Xiaochuang+Han%2C+Tianxing+He%2C+and+Yulia+Tsvetkov%2E+2024%2E+Can+LLM+graph+reasoning+generalize+beyond+pattern+memorization%3F+In+Findings+of+the+Associat)
48. Yizhuo Zhang, Heng Wang, Shangbin Feng, Zhaoxuan Tan, Xinyun Liu, and Yulia Tsvetkov. 2025. Generalizable llm learning of graph synthetic data with reinforcement learning. arXiv preprint arXiv:2506.00845. — [arXiv:2506.00845](https://arxiv.org/abs/2506.00845) [find](https://scholar.google.com/scholar?q=Yizhuo+Zhang%2C+Heng+Wang%2C+Shangbin+Feng%2C+Zhaoxuan+Tan%2C+Xinyun+Liu%2C+and+Yulia+Tsvetkov%2E+2025%2E+Generalizable+llm+learning+of+graph+synthetic+data+with+reinforcement+learning%2E+arXiv+preprint+arXiv%3A2506%2E00)
49. Zhanqiu Zhang, Jie Wang, Jiajun Chen, Shuiwang Ji, and Feng Wu. 2021. Cone: Cone embeddings for multi-hop reasoning over knowledge graphs. Advances in Neural Information Processing Systems, 34:19172–19183. [find](https://scholar.google.com/scholar?q=Zhanqiu+Zhang%2C+Jie+Wang%2C+Jiajun+Chen%2C+Shuiwang+Ji%2C+and+Feng+Wu%2E+2021%2E+Cone%3A+Cone+embeddings+for+multi%2Dhop+reasoning+over+knowledge+graphs%2E+Advances+in+Neural+Information+Processing+Systems%2C+34%3A19172%E2%80%931)
50. Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, and 1 others. 2023. [Judging llm-as-a-judge with mt-bench and chatbot arena](https://md.research.elamur.ai/free/en/papers/zheng-2023-judging-llm-judge-mt-bench/). Advances in neural information processing systems, 36:46595–46623.
51. Denny Zhou, Nathanael Schärli, Le Hou, Jason Wei, Nathan Scales, Xuezhi Wang, Dale Schuurmans, Claire Cui, Olivier Bousquet, Quoc Le, and 1 others. 2022. [Least-to-most prompting enables complex reasoning in large language models](https://md.research.elamur.ai/free/en/papers/al-2022-least-most-prompting-enables-complex/). arXiv preprint arXiv:2205.10625.

*Updated 1 November 2026*