Yu Zheng (郑瑜)

I am currently a postdoctoral fellow at Massachusetts Institute of Technology. Previously, I completed my PhD at FIBLAB, Department of Electronic Engineering, Tsinghua University in 2024. I received my bachelor degree from the Department of Electronic Engineering, Tsinghua University in 2019.

My research interests lie at the intersection of artificial intelligence, system design and science, specifically focusing on how learning-based methods can solve combinatorial design problems in physical systems, as well as the theoretical foundations of the learning-based methods. I am broadly interested in ML for System, AI for Science, and RL and its applications in real-world complex systems.

News

May 1, 2025	One paper on reinforcement learning for expensive-to-evaluate systems is accepted by ICML 2025!
Apr 28, 2025	We are organizing the AI for Complex Network tutorial at WWW’25!
Apr 27, 2025	We are organizing the Embodied Intelligence with Large Language Models In Open City Environment: From Indoor to Outdoor workshop at ICLR’25!
Sep 11, 2023	Our paper on spatial planning with reinforcement learning is published in Nature Computational Science as a cover article!

Selected Publications

NatRevPhys

Understanding emergence in complex systems using abductive AI

Jingtao Ding*, Yu Zheng*, Fengli Xu*, Carlo Vittorio Cannistraci, Xiaowen Dong, Paolo Santi, Guido Caldarelli, Yizhou Sun, Qi R Wang, Boleslaw K Szymanski, Carlo Ratti, Trey Ideker, Jianxi Gao, Yong Li, and Deliang Chen

Nature Reviews Physics, 2025

Bib HTML PDF

@article{ding2025understanding,
  title = {Understanding emergence in complex systems using abductive AI},
  author = {Ding*, Jingtao and Zheng*, Yu and Xu*, Fengli and Cannistraci, Carlo Vittorio and Dong, Xiaowen and Santi, Paolo and Caldarelli, Guido and Sun, Yizhou and Wang, Qi R and Szymanski, Boleslaw K and Ratti, Carlo and Ideker, Trey and Gao, Jianxi and Li, Yong and Chen, Deliang},
  journal = {Nature Reviews Physics},
  year = {2025},
  doi = {10.1038/s42254-025-00895-5},
  url = {https://doi.org/10.1038/s42254-025-00895-5},
}

preprint

The Thinking Spectrum: An Emperical Study of Tunable Reasoning in LLMs through Model Merging

Xiaochong Lan, Yu Zheng, Shiteng Cao, and Yong Li

preprint, 2025

arXiv Bib PDF

@article{lan2025thinking,
  title = {The Thinking Spectrum: An Emperical Study of Tunable Reasoning in LLMs through Model Merging},
  author = {Lan, Xiaochong and Zheng, Yu and Cao, Shiteng and Li, Yong},
  journal = {preprint},
  year = {2025},
}

preprint

GeoEvolve: Automating Geospatial Model Discovery via Multi-Agent Large Language Models

Peng Luo, Xiayin Lou, Yu Zheng, Zhuo Zheng, and Stefano Ermon

preprint, 2025

arXiv Bib PDF

@article{luo2025geoevolve,
  title = {GeoEvolve: Automating Geospatial Model Discovery via Multi-Agent Large Language Models},
  author = {Luo, Peng and Lou, Xiayin and Zheng, Yu and Zheng, Zhuo and Ermon, Stefano},
  journal = {preprint},
  year = {2025},
}

preprint

Probing Neural Topology of Large Language Models

Yu Zheng, Yuan Yuan, Yue Zhuo, Yong Li, and Paolo Santi

preprint, 2025

arXiv Bib PDF Code

@article{zheng2025llm,
  title = {Probing Neural Topology of Large Language Models},
  author = {Zheng, Yu and Yuan, Yuan and Zhuo, Yue and Li, Yong and Santi, Paolo},
  journal = {preprint},
  year = {2025},
}

preprint

Advancing Network Resilience Theories with Symbolized Reinforcement Learning

Yu Zheng, Jingtao Ding, Depeng Jin, Jianxi Gao, and Yong Li

preprint, 2025

arXiv Bib PDF Supp Code

@article{zheng2025selinda,
  title = {Advancing Network Resilience Theories with Symbolized Reinforcement Learning},
  author = {Zheng, Yu and Ding, Jingtao and Jin, Depeng and Gao, Jianxi and Li, Yong},
  journal = {preprint},
  year = {2025},
}

ICML

Reinforcement Learning with Adaptive Reward Modeling for Expensive-to-Evaluate Systems

Hongyuan Su*, Yu Zheng*, Yuan Yuan, Yuming Lin, Depeng Jin, and Yong Li

In International Conference on Machine Learning, 2025

Bib PDF Code

@inproceedings{su2025adaremo,
  title = {Reinforcement Learning with Adaptive Reward Modeling for Expensive-to-Evaluate Systems},
  author = {Su*, Hongyuan and Zheng*, Yu and Yuan, Yuan and Lin, Yuming and Jin, Depeng and Li, Yong},
  booktitle = {International Conference on Machine Learning},
  year = {2025},
  organization = {PMLR},
  address = {Vancouver, Canada},
}

NatComputSci

Spatial planning of urban communities via deep reinforcement learning

Yu Zheng, Yuming Lin, Liang Zhao, Tinghai Wu, Depeng Jin, and Yong Li

Nature Computational Science, 2023

Bib HTML PDF Supp Code Poster

@article{zheng2023spatial,
  title = {Spatial planning of urban communities via deep reinforcement learning},
  author = {Zheng, Yu and Lin, Yuming and Zhao, Liang and Wu, Tinghai and Jin, Depeng and Li, Yong},
  journal = {Nature Computational Science},
  pages = {1--15},
  year = {2023},
  publisher = {Nature Publishing Group US New York},
  doi = {10.1038/s43588-023-00503-5},
  url = {https://www.nature.com/articles/s43588-023-00503-5},
}