Elamur Research⌕
☰
home / papers / 2026 / EvoMD-LLM: Learning the Language of Species Evolution in Reactive Molecular Dynamics

EvoMD-LLM: Learning the Language of Species Evolution in Reactive Molecular Dynamics

Paper · 2026 · Zhichen Tang, Zhengzheng Dang, Yulin Chen et al.

Summary

LLMs handle static chemistry well, but they have trouble capturing how reactive systems change over time. EvoMD-LLM addresses this by recasting reactive molecular dynamics (MD) as a symbolic sequence task. Raw trajectories are converted into a sequence of species events:

  • an ab initio bond-order criterion identifies which atoms are bonded in each frame;
  • connected components of the resulting graph are turned into molecular formulas;
  • each formula is paired with the time it persists.

Each formula is kept as one indivisible token rather than split into characters. A duration token is placed next to every species, an idea the authors call temporal scaffolding and compare to run-length encoding. The duration acts as a signal of kinetic stability. An instruction-tuned autoregressive LLM is then fine-tuned on sliding windows of these events. The same model handles several prediction tasks, including forecasting forward and inferring backward.

On a Mo-S chemical vapor deposition (CVD) system, the model reaches up to 66.14% accuracy and beats both sequential neural networks and language-based baselines. Duration tokens noticeably cut down on invalid or hallucinated species. The model also produces chemically plausible explanations for its predictions, even though it was never trained on such explanations. This suggests LLMs can serve as general learners of time-resolved simulation data.

Key points

  • Events are band-pass filtered to durations between 10 and 500 ps. Each training example uses a context of 3–5 past events to predict the next one.
  • Two-dimensional stratified sampling over species identity and duration bins (short, medium, long) counters the long-tail species distribution.
  • Ablations show that temporal scaffolding raises accuracy and reduces outputs of invalid molecules. The best reported accuracy is 66.14%.
  • Qwen3 was used as an automated judge to score the qualitative explanations. Example cases cover Mo–O/S species such as MoS, MoS₃ and MoS₅.
  • Stated limitations:
    • evaluation covers only the Mo-S system;
    • errors accumulate in multi-step (N-step) prediction;
    • the symbolic representation loses fine-grained geometry;
    • explanations can be plausible yet ungrounded.

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

🎓Shanghai Jiao Tong UniversityShanghai · China212 works · 19,544 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Physics and Astronomy · 52 worksStatistical and Nonlinear Physics · 47 worksMaterials Science · 31 worksMaterials Chemistry · 31 worksChemistry · 4 worksPhysical and Theoretical Chemistry · 4 works

🧭 Topics 1Research subjects and languages the work is about.Language models & pretraining · 3,894 works

Cite

Cite as Zhichen Tang, Zhengzheng Dang, Yulin Chen, Jixin Wu, Haiwen Li and Yanming Wang. 2026. EvoMD-LLM: Learning the Language of Species Evolution in Reactive Molecular Dynamics. 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.1947

@inproceedings{tang2026evomd,
  title = {{EvoMD-LLM: Learning the Language of Species Evolution in Reactive Molecular Dynamics}},
  author = {Zhichen Tang, Zhengzheng Dang, Yulin Chen, Jixin Wu, Haiwen Li and Yanming Wang},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1947},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1947}
}

References

35 references, 13 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D. Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, Sandhini Agarwal, Ariel Herbert-Voss, Gretchen Krueger, Tom Henighan, Rewon Child, Aditya Ramesh, Daniel M. Ziegler, Jeffrey Wu, Clemens Winter,… find
  2. J. Cavanagh, K. Sun, A. Gritsevskiy, D. Bagni, T. Head-Gordon, and T. D. Bannister. 2024. Smileyllama: Modifying large language models for directed chemical space exploration. In NeurIPS 2024 Workshop on AI for New Drug Modalities. find
  3. Xiuying Chen, Tairan Wang, Taicheng Guo, Kehan Guo, Juexiao Zhou, Haoyang Li, Zirui Song, Xin Gao, and Xiangliang Zhang. 2025. Unveiling the power of language models in chemical research question answering. Communications Chemistry, 8(1):4.
  4. Seyone Chithrananda, Gabriel Grand, and Bharath Ramsundar. 2020. Chemberta: Large-scale self-supervised pretraining for molecular property prediction. CoRR, abs/2010.09885. find
  5. Peter Clark, Isaac Cowhey, Oren Etzioni, Tushar Khot, Ashish Sabharwal, Carissa Schoenick, and Oyvind Tafjord. 2018. Think you have solved question answering? try arc, the ai2 reasoning challenge. CoRR, abs/1803.05457.
  6. Rishabh Agarwal, Avi Singh, Lei M. Zhang, Bernd Bohnet, Luis Rosias, Stephanie C. Y. Chan, Biao Zhang, Ankesh Anand, Zaheer Abbas, Azade Nova, John D. Co-Reyes, Eric Chu, Feryal M. P. Behbahani, Aleksandra Faust, and Hugo Larochelle. 2024. Many-shot in-context learning. Advances in Neural Informatio…
  7. Connor W Coley, Wengong Jin, Luke Rogers, Timothy F Jamison, Tommi S Jaakkola, William H Green, Regina Barzilay, and Klavs F Jensen. 2019. A graph-convolutional neural network model for the prediction of chemical reactivity. Chemical science, 10(2):370–377. find
  8. Berni Julian Alder and Thomas Everett Wainwright. 1957. Phase transition for a hard sphere system. The Journal of chemical physics, 27(5):1208. find
  9. OpenCompass Contributors. 2023. Opencompass: A universal evaluation platform for foundation models. https://github.com/open-compass/ opencompass. ↗
  10. Abdul Fatir Ansari, Lorenzo Stella, Caner Turkmen, Xiyuan Zhang, Pedro Mercado, Huibin Shen, Oleksandr Shchur, Syama Sundar Rangapuram, Sebastian Pineda Arango, Shubham Kapoor, Jasper Zschiegner, Danielle C. Maddix, Hao Wang, Michael W. Mahoney, Kari Torkkola, Andrew Gordon Wilson, Michael Bohlke-Sc… find
  11. Palash Bera and Jagannath Mondal. 2025. Accurate prediction of the kinetic sequence of physicochemical states using generative artificial intelligence. Chemical Science, 16(20):8735–8751. find
  12. Daniil A Boiko, Robert MacKnight, and Gabe Gomes. 2023. Emergent autonomous scientific research capabilities of large language models. arXiv preprint arXiv:2304.05332.
  13. Zhengzheng Dang, Zhichen Tang, Jixin Wu, Yide Chang, and Yanming Wang. 2025. Unraveling the reaction networks and key pathways during the gas phase stage in cvd synthesis of mos2. Chemical Engineering Journal, 503:157957. find
  14. Qingxiu Dong, Lei Li, Damai Dai, Ce Zheng, Jingyuan Ma, Rui Li, Heming Xia, Jingjing Xu, Zhiyong Wu, Baobao Chang, Xu Sun, and Zhifang Sui. 2024. A survey on in-context learning. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 1107–1128.
  15. Daniel Han, Michael Han, and Unsloth team. 2023. Unsloth. find
  16. S. Hochreiter and J. Schmidhuber. 1997. Long short-term memory. Neural Computation, 9(8):1735– 1780. find
  17. 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. ICLR, 1(2):3.
  18. Cheng-Zhi Anna Huang, Ashish Vaswani, Jakob Uszkoreit, Ian Simon, Curtis Hawthorne, Noam Shazeer, Andrew M Dai, Matthew D Hoffman, Monica Dinculescu, and Douglas Eck. 2018. Music transformer: Generating music with long-term structure. In International Conference on Learning Representations. find
  19. Mhd Hussein Murtada, Z Faidon Brotzakis, and Michele Vendruscolo. 2025. Md-llm-1: A large language model for molecular dynamics. arXiv e-prints, pages arXiv–2508. find
  20. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Sebastian Riedel, and Douwe Kiela. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in neural information process…
  21. Ilya Loshchilov and Frank Hutter. 2019. Decoupled weight decay regularization. In International Conference on Learning Representations. find
  22. Feifei Luo, Jinglang Zhang, Qilong Wang, and Chunpeng Yang. 2025. Leveraging prompt engineering in large language models for accelerating chemical research. ACS Central Science, 11(4):511–519. find
  23. Meta AI Team. 2024. The llama 3 herd of models: the llama 3.1 family (405b parameters, 128k context window). Technical report. Technical Report.
  24. Mhd Hussein Murtada, Z Faidon Brotzakis, and Michele Vendruscolo. 2024. Language models for molecular dynamics. bioRxiv, pages 2024–11. find
  25. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll L. Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, John Schulman, Jacob Hilton, Fraser Kelton, Luke Miller, Maddie Simens, Amanda Askell, Peter Welinder, Paul F. Christiano, Jan Leike, and Ryan Lowe. 2022.… — Training language models to follow instructions with human feedback
  26. Yi Ren, Yangjun Ruan, Xu Tan, Tao Qin, Sheng Zhao, Zhou Zhao, and Tie-Yan Liu. 2019. Fastspeech: Fast, robust and controllable text to speech. Advances in neural information processing systems, 32. find
  27. Khalid Sayood. 2017. Introduction to Data Compression, 5th edition. Morgan Kaufmann, Burlington, MA, USA. find
  28. Rico Sennrich, Barry Haddow, and Alexandra Birch. 2016. Neural machine translation of rare words with subword units. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1715–1725, Berlin, Germany. Association for Computational Lin…
  29. 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. find
  30. 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
  31. Sun-Ting Tsai, En-Jui Kuo, and Pratyush Tiwary. 2020. Learning molecular dynamics with simple language model built upon long short-term memory neural network. Nature communications, 11(1):5115. find
  32. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in neural information processing systems, 30.
  33. Daniel S Wigh, Jonathan M Goodman, and Alexei A Lapkin. 2022. A review of molecular representation in the age of machine learning. Wiley Interdisciplinary Reviews: Computational Molecular Science, 12(5):e1603. find
  34. A. Yang, A. Li, B. Yang, B. Zhang, B. Hui, B. Zheng, and Z. Qiu. 2025. Qwen3 technical report. arXiv preprint arXiv:2505.09388.
  35. Xianrui Zhong, Bowen Jin, Siru Ouyang, Yanzhen Shen, Qiao Jin, Yin Fang, Zhiyong Lu, and Jiawei Han. 2025. Benchmarking retrieval-augmented generation for chemistry. In Second Conference on Language Modeling. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/tang-2026-evomd-llm-learning-language-species/