Elamur Research⌕
☰
home / papers / 2026 / Temp-R1: A Unified Autonomous Agent for Complex Temporal KGQA via Reverse Curriculum Reinforcement Learning

Temp-R1: A Unified Autonomous Agent for Complex Temporal KGQA via Reverse Curriculum Reinforcement Learning

Paper · 2026 · Zhaoyan Gong, Zhiqiang Liu, Songze Li et al.

Summary

Temp-R1 addresses question answering over temporal knowledge graphs (TKGQA), where facts are timestamped quadruples and questions often combine multi-hop links with time constraints. Earlier LLM-based systems use hand-designed pipelines with separate planners, retrievers and generators. They also depend on costly closed-source APIs. The authors instead train a single 8B open model as an autonomous agent using reinforcement learning (GRPO), after a short SFT cold start. Search-R1-style agents put all reasoning in one <think> block. Temp-R1 instead splits it into separate internal actions, <plan>, <filter> and <rank>, alongside the external <search> call to the TKG and a final <answer>. To keep the agent from relying on shortcuts that only work for easy questions, training follows a reverse curriculum: the hardest questions come first, then the easier ones. The agent reaches state-of-the-art results on MultiTQ and TimelineKGQA and beats GPT-4o-based methods. This shows that a small, cheap model can learn flexible temporal reasoning without a fixed workflow.

Key points

  • Action space: an external <search> action over the TKG, plus internal <plan>, <filter> (semantic and temporal constraints) and <rank> (chronological ordering) actions. These separate steps aim to reduce the hallucinations caused by overloading a single reasoning step.
  • Training pipeline: SFT warmup, then GRPO with a hard-first curriculum. Without SFT, training collapsed and KL loss spiked. Easy-first ordering plateaued early.
  • On complex "Multiple"-type questions, Temp-R1 improves by 19.8% over the strongest baseline. It is evaluated on three datasets, including MultiTQ and TimelineKGQA.
  • Backbones: peak accuracy was 0.790 for the 7B Qwen model versus 0.532 for 1.5B, and Llama reached 0.780. On complex questions the agent used more steps (<think> rose from 1.36 to 2.93, <search> from 1.33 to 1.92).
  • Code is released at github.com/zjukg/Temp-R1. The authors note two limitations: experiments go only up to 8B, and the generality of reverse curriculum outside TKGQA has not been tested.

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

🎓Zhejiang UniversityHangzhou · China172 works · 12,988 citations🏢Ant GroupHangzhou · China42 works · 1,131 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

Cite

Cite as Zhaoyan Gong, Zhiqiang Liu, Songze Li, Xiaoke Guo, Yuanxiang Liu, Xinle Deng, Zhizhen Liu, Lei Liang, Huajun Chen and Wen Zhang. 2026. Temp-R1: A Unified Autonomous Agent for Complex Temporal KGQA via Reverse Curriculum Reinforcement Learning. 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. DOI 10.18653/v1/2026.acl-long.1184

@inproceedings{gong2026temp,
  title = {{Temp-R1: A Unified Autonomous Agent for Complex Temporal KGQA via Reverse Curriculum Reinforcement Learning}},
  author = {Zhaoyan Gong, Zhiqiang Liu, Songze Li, Xiaoke Guo, Yuanxiang Liu, Xinle Deng, Zhizhen Liu, Lei Liang, Huajun Chen and Wen Zhang},
  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.1184},
  url = {https://doi.org/10.18653/v1/2026.acl-long.1184}
}

References

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

  1. Huajun Chen. 2024. Large knowledge model: Perspectives and challenges. DATA INTELLIGENCE, 6(3):587–620. find
  2. Mingyang Chen, Tianpeng Li, Haoze Sun, Yijie Zhou, Chenzheng Zhu, Haofen Wang, Jeff Z. Pan, Wen Zhang, Huajun Chen, Fan Yang, Zenan Zhou, and Weipeng Chen. 2025. Research: Learning to reason with search for llms via reinforcement learning. CoRR, abs/2503.19470. find
  3. Zhuo Chen, Zhao Zhang, Zixuan Li, Fei Wang, Yutao Zeng, Xiaolong Jin, and Yongjun Xu. 2024a. Self-improvement programming for temporal knowledge graph question answering. In LREC/COLING, pages 14579–14594. ELRA and ICCL. find
  4. Ziyang Chen, Dongfang Li, Xiang Zhao, Baotian Hu, and Min Zhang. 2024b. Temporal knowledge question answering via abstract reasoning induction. In ACL (1), pages 4872–4889. Association for Computational Linguistics. find
  5. Ziyang Chen, Jinzhi Liao, and Xiang Zhao. 2023. Multi-granularity temporal question answering over knowledge graphs. In ACL (1), pages 11378–11392. Association for Computational Linguistics. find
  6. Ziyang Chen, Xiang Zhao, Jinzhi Liao, Xinyi Li, and Evangelos Kanoulas. 2022. Temporal knowledge graph question answering via subgraph reasoning. Knowl. Based Syst., 251:109134. find
  7. Wentao Ding, Hao Chen, Huayu Li, and Yuzhong Qu. 2022. Semantic framework based query generation for temporal question answering over knowledge graphs. In EMNLP, pages 1867–1877. Association for Computational Linguistics. find
  8. Yifu Gao, Linbo Qiao, Zhigang Kan, Zhihua Wen, Yongquan He, and Dongsheng Li. 2024. Two-stage generative question answering on temporal knowledge graph using large language models. In ACL (Findings), pages 6719–6734. Association for Computational Linguistics. find
  9. Zhaoyan Gong, Juan Li, Zhiqiang Liu, Lei Liang, Huajun Chen, and Wen Zhang. 2025. RTQA : Recursive thinking for complex temporal knowledge graph question answering with large language models. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 9864–9881,…
  10. Daya Guo, Dejian Yang, Haowei Zhang, Junxiao Song, Peiyi Wang, Qihao Zhu, Runxin Xu, Ruoyu Zhang, Shirong Ma, Xiao Bi, and 1 others. 2025. Deepseekr1 incentivizes reasoning in llms through reinforcement learning. Nature, 645(8081):633–638. find
  11. Xiaoke Guo, Songze Li, Zhiqiang Liu, Zhaoyan Gong, Yuanxiang Liu, Huajun Chen, and Wen Zhang. 2026. Astra: Adaptive semantic tree reasoning architecture for complex table question answering. Preprint, arXiv:2604.08999. find
  12. Qianyi Hu, Xinhui Tu, Cong Guo, and Shunping Zhang. 2025. Time-aware ReAct agent for temporal knowledge graph question answering. In Findings of the Association for Computational Linguistics: NAACL 2025, pages 6013–6024, Albuquerque, New Mexico. Association for Computational Linguistics. find
  13. Zhen Jia, Abdalghani Abujabal, Rishiraj Saha Roy, Jannik Strötgen, and Gerhard Weikum. 2018a. Tempquestions: A benchmark for temporal question answering. In WWW (Companion Volume), pages 1057– 1062. ACM. find
  14. Zhen Jia, Abdalghani Abujabal, Rishiraj Saha Roy, Jannik Strötgen, and Gerhard Weikum. 2018b. TEQUILA: temporal question answering over knowledge bases. In CIKM, pages 1807–1810. ACM. find
  15. Zhen Jia, Philipp Christmann, and Gerhard Weikum. 2024. Faithful temporal question answering over heterogeneous sources. In WWW, pages 2052–2063. ACM. find
  16. Zhen Jia, Soumajit Pramanik, Rishiraj Saha Roy, and Gerhard Weikum. 2021. Complex temporal question answering on knowledge graphs. In CIKM, pages 792–802. ACM. find
  17. Bowen Jin, Hansi Zeng, Zhenrui Yue, Jinsung Yoon, Sercan Arik, Dong Wang, Hamed Zamani, and Jiawei Han. 2025. Search-r1: Training llms to reason and leverage search engines with reinforcement learning. arXiv preprint arXiv:2503.09516.
  18. Leslie Pack Kaelbling, Michael L. Littman, and Andrew W. Moore. 1996. Reinforcement learning: A survey. J. Artif. Intell. Res., 4:237–285. find
  19. Songze Li, Xiaoke Guo, Tianqi Liu, Biao Yi, Zhaoyan Gong, Zhiqiang Liu, Huajun Chen, and Wen Zhang. 2026. What’s missing in screen-to-action? towards a ui-in-the-loop paradigm for multimodal gui reasoning. Preprint, arXiv:2604.06995. find
  20. Yonghao Liu, Di Liang, Fang Fang, Sirui Wang, Wei Wu, and Rui Jiang. 2023a. Time-aware multiway adaptive fusion network for temporal knowledge graph question answering. In ICASSP, pages 1–5. IEEE. find
  21. Yonghao Liu, Di Liang, Mengyu Li, Fausto Giunchiglia, Ximing Li, Sirui Wang, Wei Wu, Lan Huang, Xiaoyue Feng, and Renchu Guan. 2023b. Local and global: Temporal question answering via information fusion. In IJCAI, pages 5141–5149. ijcai.org. find
  22. Yuanxiang Liu, Songze Li, Xiaoke Guo, Zhaoyan Gong, Qifei Zhang, Huajun Chen, and Wen Zhang. 2026. Cog: Controllable graph reasoning via relational blueprints and failure-aware refinement over knowledge graphs. Preprint, arXiv:2601.11047. find
  23. Zhiqiang Liu, Enpei Niu, Yin Hua, Mengshu Sun, Lei Liang, Huajun Chen, and Wen Zhang. 2025a. SKA-bench: A fine-grained benchmark for evaluating structured knowledge understanding of LLMs. In Findings of the Association for Computational Linguistics: EMNLP 2025, pages 3626–3640, Suzhou, China. Associ… find
  24. Zhiqiang Liu, Yichi Zhang, Mengshu Sun, Lei Liang, and Wen Zhang. 2025b. Collaboration of fusion and independence: Hypercomplex-driven robust multi-modal knowledge graph completion. Preprint, arXiv:2509.23714. find
  25. Xinji Mai, Haotian Xu, Xing W, Weinong Wang, Yingying Zhang, and Wenqiang Zhang. 2025. Agent RL scaling law: Agent RL with spontaneous code execution for mathematical problem solving. CoRR, abs/2505.07773. — [arXiv:2505.07773](https://arxiv.org/abs/2505.07773) find
  26. Costas Mavromatis, Prasanna Lakkur Subramanyam, Vassilis N. Ioannidis, Adesoji Adeshina, Phillip Ryan Howard, Tetiana Grinberg, Nagib Hakim, and George Karypis. 2022. Tempoqr: Temporal question reasoning over knowledge graphs. In AAAI, pages 5825– 5833. AAAI Press. find
  27. Songze Li, Zhiqiang Liu, Zhaoyan Gong, Xiaoke Guo, Zhengke Gui, Huajun Chen, and Wen Zhang. 2025a. Last layer logits to logic: Empowering llms with logic-consistent structured knowledge reasoning. Preprint, arXiv:2511.07910. — [arXiv:2511.07910](https://arxiv.org/abs/2511.07910) find
  28. Sumit Neelam, Udit Sharma, Hima Karanam, Shajith Ikbal, Pavan Kapanipathi, Ibrahim Abdelaziz, Nandana Mihindukulasooriya, Young-Suk Lee, Santosh K. Srivastava, Cezar Pendus, Saswati Dana, Dinesh Garg, Achille Fokoue, G. P. Shrivatsa Bhargav, Dinesh Khandelwal, Srinivas Ravishankar, Sairam Gurajada,… — [arXiv:2109.13430](https://arxiv.org/abs/2109.13430) find
  29. Songze Li, Zhiqiang Liu, Zhengke Gui, Huajun Chen, and Wen Zhang. 2025b. Enrich-on-graph: Query-graph alignment for complex reasoning with LLM enriching. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 7672–7692, Suzhou, China. Association for Computa…
  30. OpenAI, Aaron Jaech, Adam Kalai, Adam Lerer, Adam Richardson, Ahmed El-Kishky, Aiden Low, Alec Helyar, Aleksander Madry, Alex Beutel, Alex Carney, Alex Iftimie, Alex Karpenko, Alex Tachard Passos, Alexander Neitz, Alexander Prokofiev, Alexander Wei, Allison Tam, Ally Bennett, and 243 others. 2024. O… — OpenAI o1 System Card
  31. Xinying Qian, Ying Zhang, Yu Zhao, Baohang Zhou, Xuhui Sui, and Xiaojie Yuan. 2025. Plan of knowledge: Retrieval-augmented large language models for temporal knowledge graph question answering. Preprint, arXiv:2511.04072. — [arXiv:2511.04072](https://arxiv.org/abs/2511.04072) find
  32. Xinying Qian, Ying Zhang, Yu Zhao, Baohang Zhou, Xuhui Sui, Li Zhang, and Kehui Song. 2024. Timer4 : Time-aware retrieval-augmented large language models for temporal knowledge graph question answering. In EMNLP, pages 6942–6952. Association for Computational Linguistics. find
  33. Rafael Rafailov, Archit Sharma, Eric Mitchell, Christopher D. Manning, Stefano Ermon, and Chelsea Finn. 2023. Direct preference optimization: Your language model is secretly a reward model. In Advances in Neural Information Processing Systems 36: Annual Conference on Neural Information Processing Sy…
  34. Apoorv Saxena, Soumen Chakrabarti, and Partha P. Talukdar. 2021. Question answering over temporal knowledge graphs. In ACL/IJCNLP (1), pages 6663– 6676. Association for Computational Linguistics. find
  35. Jing Wang, Shuo Zhang, and Runzhi Li. 2024. Gate feature interaction network for relation prediction in knowledge graph. DATA INTELLIGENCE, 6(3):749– 770. find
  36. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik R. Narasimhan, and Yuan Cao. 2023. React: Synergizing reasoning and acting in language models. In ICLR. OpenReview.net.
  37. Yichi Zhang, Zhuo Chen, Lingbing Guo, Lei Liang, Wen Zhang, and Huajun Chen. 2025. Structured and abstractive reasoning on multi-modal relational knowledge images. Preprint, arXiv:2510.21828.
  38. Chujie Zheng, Shixuan Liu, Mingze Li, Xiong-Hui Chen, Bowen Yu, Chang Gao, Kai Dang, Yuqiong Liu, Rui Men, An Yang, Jingren Zhou, and Junyang Lin. 2025. Group sequence policy optimization. CoRR, abs/2507.18071.
  39. Jingsheng Zheng, Jintian Zhang, Yujie Luo, Yuren Mao, Yunjun Gao, Lun Du, Huajun Chen, and Ningyu Zhang. 2026. Can we predict before executing machine learning agents? Preprint, arXiv:2601.05930.
  40. Apoorv Saxena, Aditay Tripathi, and Partha P. Talukdar. 2020. Improving multi-hop question answering over knowledge graphs using knowledge base embeddings. In ACL, pages 4498–4507. Association for Computational Linguistics.
  41. John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. 2017. Proximal policy optimization algorithms. CoRR, abs/1707.06347. find
  42. Aditya Sharma, Apoorv Saxena, Chitrank Gupta, Seyed Mehran Kazemi, Partha P. Talukdar, and Soumen Chakrabarti. 2023. Twirgcn: Temporally weighted graph convolution for question answering over temporal knowledge graphs. In EACL, pages 2041–2052. Association for Computational Linguistics. find
  43. Joykirat Singh, Raghav Magazine, Yash Pandya, and Akshay Nambi. 2025. Agentic reasoning and tool integration for llms via reinforcement learning. CoRR, abs/2505.01441. — [arXiv:2505.01441](https://arxiv.org/abs/2505.01441) find
  44. Huatong Song, Jinhao Jiang, Yingqian Min, Jie Chen, Zhipeng Chen, Wayne Xin Zhao, Lei Fang, and Ji-Rong Wen. 2025. R1-searcher: Incentivizing the search capability in llms via reinforcement learning. CoRR, abs/2503.05592.
  45. Miao Su, Zixuan Li, Zhuo Chen, Long Bai, Xiaolong Jin, and Jiafeng Guo. 2024. Temporal knowledge graph question answering: A survey. CoRR, abs/2406.14191. — [arXiv:2406.14191](https://arxiv.org/abs/2406.14191) find
  46. Xingyu Tan, Xiaoyang Wang, Qing Liu, Xiwei Xu, Xin Yuan, Liming Zhu, and Wenjie Zhang. 2025. Memotime: Memory-augmented temporal knowledge graph enhanced large language model reasoning. Preprint, arXiv:2510.13614. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/gong-2026-temp-r1-unified-autonomous-agent/