Publications & Projects

For the updated list, check out my Google Scholar page.

(* denotes equal contribution or alphabetic ordering, † denotes corresponding author)

ScaleAutoResearch: Scaling Automated Research for Scientific Discovery

Yiping Wang

A self-built system that runs many autonomous frontier-LLM coding agents in parallel to scale automated research, driving results such as new Ramsey-number lower bounds and nanoGPT-speedrun optimizer discovery.

Project

Code (Ramsey) ★ 47 X (Ramsey)量子位 (Ramsey)X (nanoGPT)X (nanoGPT update)

BibTeX

@misc{wang2026scaleautoresearchramsey,
  author       = {Yiping Wang},
  title        = {ScaleAutoResearch-Ramsey},
  howpublished = {https://github.com/ypwang61/ScaleAutoResearch-Ramsey},
  note         = {Verified graph witnesses and experiment records for new lower bounds on classical Ramsey numbers},
  year         = {2026}
}

Visored: A Controlled-Natural-Language Prover for LLM-Generated Mathematics

Xiyu Zhai, Xinyi Chen, Yiping Wang, Runlong Zhou, Liao Zhang, Simon S. Du

Preprint 2026

Arxiv

BibTeX

@misc{zhai2026visoredcontrollednaturallanguageproverllmgenerated,
      title={Visored: A Controlled-Natural-Language Prover for LLM-Generated Mathematics}, 
      author={Xiyu Zhai and Xinyi Chen and Yiping Wang and Runlong Zhou and Liao Zhang and Simon S. Du},
      year={2026},
      eprint={2606.17581},
      archivePrefix={arXiv},
      primaryClass={cs.PL},
      url={https://arxiv.org/abs/2606.17581}, 
}

Latent Recurrent Transformer: Architecture Exploration, Training Strategies, and Scaling Behavior

Zeyi Huang, Xuehai He, Liliang Ren, Yiping Wang, Baolin Peng, Hao Cheng, Shuohang Wang, Pengcheng He, Jianfeng Gao, Yong Jae Lee, Yelong Shen

Preprint 2026

Arxiv

BibTeX

@misc{huang2026latentrecurrenttransformerarchitecture,
      title={Latent Recurrent Transformer: Architecture Exploration, Training Strategies, and Scaling Behavior}, 
      author={Zeyi Huang and Xuehai He and LiLiang Ren and Yiping Wang and Baolin Peng and Hao Cheng and Shuohang Wang and Pengcheng He and Jianfeng Gao and Yong Jae Lee and Yelong Shen},
      year={2026},
      eprint={2605.26797},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2605.26797}, 
}

ThetaEvolve: Test-time Learning on Open Problems

Yiping Wang†, Shao-Rong Su, Zhiyuan Zeng, Eva Xu, Liliang Ren, Xinyu Yang, Zeyi Huang, Xuehai He, Luyao Ma, Baolin Peng, Hao Cheng, Pengcheng He, Weizhu Chen, Shuohang Wang, Simon Shaolei Du†, Yelong Shen†

ICML 2026

Arxiv Code ★ 164 X

BibTeX

@misc{wang2025thetaevolvetesttimelearningopen,
      title={ThetaEvolve: Test-time Learning on Open Problems}, 
      author={Yiping Wang and Shao-Rong Su and Zhiyuan Zeng and Eva Xu and Liliang Ren and Xinyu Yang and Zeyi Huang and Xuehai He and Luyao Ma and Baolin Peng and Hao Cheng and Pengcheng He and Weizhu Chen and Shuohang Wang and Simon Shaolei Du and Yelong Shen},
      year={2025},
      eprint={2511.23473},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2511.23473}, 
}

RLVE: Scaling Up Reinforcement Learning for Language Models with Adaptive Verifiable Environments

Zhiyuan Zeng*, Hamish Ivison*, Yiping Wang*, Lifan Yuan*, Shuyue Stella Li, Zhuorui Ye, Siting Li, Jacqueline He, Runlong Zhou, Tong Chen, Chenyang Zhao, Yulia Tsvetkov, Simon Shaolei Du, Natasha Jaques, Hao Peng, Pang Wei Koh, Hannaneh Hajishirzi

ICML 2026

Arxiv Code ★ 222 X

BibTeX

@misc{zeng2026rlvescalingreinforcementlearning,
      title={RLVE: Scaling Up Reinforcement Learning for Language Models with Adaptive Verifiable Environments}, 
      author={Zhiyuan Zeng and Hamish Ivison and Yiping Wang and Lifan Yuan and Shuyue Stella Li and Zhuorui Ye and Siting Li and Jacqueline He and Runlong Zhou and Tong Chen and Chenyang Zhao and Yulia Tsvetkov and Simon Shaolei Du and Natasha Jaques and Hao Peng and Pang Wei Koh and Hannaneh Hajishirzi},
      year={2026},
      eprint={2511.07317},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2511.07317}, 
}

Spurious Rewards: Rethinking Training Signals in RLVR

Rulin Shao*, Shuyue Stella Li*, Rui Xin*, Scott Geng*, Yiping Wang, Sewoong Oh, Simon Shaolei Du, Nathan Lambert, Sewon Min, Ranjay Krishna, Yulia Tsvetkov, Hannaneh Hajishirzi, Pang Wei Koh, Luke Zettlemoyer

ICML 2026

Arxiv Blog1 Blog2 Code ★ 358 W&B Models X

BibTeX

@misc{shao2026spuriousrewardsrethinkingtraining,
      title={Spurious Rewards: Rethinking Training Signals in RLVR}, 
      author={Rulin Shao and Shuyue Stella Li and Rui Xin and Scott Geng and Yiping Wang and Sewoong Oh and Simon Shaolei Du and Nathan Lambert and Sewon Min and Ranjay Krishna and Yulia Tsvetkov and Hannaneh Hajishirzi and Pang Wei Koh and Luke Zettlemoyer},
      year={2026},
      eprint={2506.10947},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2506.10947}, 
}

Reinforcement Learning for Reasoning in Large Language Models with One Training Example

Yiping Wang†, Qing Yang, Zhiyuan Zeng, Liliang Ren, Lucas Liu, Baolin Peng, Hao Cheng, Xuehai He, Kuan Wang, Jianfeng Gao, Weizhu Chen, Shuohang Wang†, Simon Shaolei Du†, Yelong Shen†

#1 Paper of the day on HuggingFace Daily Papers

We show that for RLVR on LLMs, one proper training example can already bring non-trivial improvement.

NeurIPS 2025

Arxiv Code ★ 441 BAAI Talk W&B Models X Slides Poster

BibTeX

@misc{wang2025reinforcementlearningreasoninglarge,
      title={Reinforcement Learning for Reasoning in Large Language Models with One Training Example}, 
      author={Yiping Wang and Qing Yang and Zhiyuan Zeng and Liliang Ren and Liyuan Liu and Baolin Peng and Hao Cheng and Xuehai He and Kuan Wang and Jianfeng Gao and Weizhu Chen and Shuohang Wang and Simon Shaolei Du and Yelong Shen},
      year={2025},
      eprint={2504.20571},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2504.20571}, 
}

FloE: On-the-Fly MoE Inference on Memory-constrained GPU

Yuxin Zhou*, Zheng Li*, Jun Zhang, Jue Wang, Yiping Wang, Zhongle Xie, Ke Chen, Lidan Shou

ICML 2025

Arxiv

BibTeX

@misc{zhou2025floeontheflymoeinference,
      title={FloE: On-the-Fly MoE Inference on Memory-constrained GPU}, 
      author={Yuxin Zhou and Zheng Li and Jun Zhang and Jue Wang and Yiping Wang and Zhongle Xie and Ke Chen and Lidan Shou},
      year={2025},
      eprint={2505.05950},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2505.05950}, 
}

SHARP: Accelerating Language Model Inference by SHaring Adjacent layers with Recovery Parameters

Yiping Wang, Hanxian Huang, Yifang Chen, Jishen Zhao, Simon Shaolei Du, Yuandong Tian

preprint 2025

Arxiv

BibTeX

@misc{wang2025sharpacceleratinglanguagemodel,
      title={SHARP: Accelerating Language Model Inference by SHaring Adjacent layers with Recovery Parameters}, 
      author={Yiping Wang and Hanxian Huang and Yifang Chen and Jishen Zhao and Simon Shaolei Du and Yuandong Tian},
      year={2025},
      eprint={2502.07832},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2502.07832}, 
}

Is Your World Simulator a Good Story Presenter? A Consecutive Events-Based Benchmark for Future Long Video Generation

Yiping Wang, Xuehai He, Kuan Wang, Luyao Ma, Jianwei Yang, Shuohang Wang, Simon Shaolei Du, Yelong Shen

Current top video generative models can not present multi-event stories like "How to Put an Elephant in a Refrigerator".

CVPR 2025

Arxiv Code Poster X Website

BibTeX

@misc{wang2024worldsimulatorgoodstory,
      title={Is Your World Simulator a Good Story Presenter? A Consecutive Events-Based Benchmark for Future Long Video Generation}, 
      author={Yiping Wang and Xuehai He and Kuan Wang and Luyao Ma and Jianwei Yang and Shuohang Wang and Simon Shaolei Du and Yelong Shen},
      year={2024},
      eprint={2412.16211},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2412.16211}, 
}

Mojito: Motion trajectory and intensity control for video generation

Xuehai He, Shuohang Wang, Jianwei Yang, Xiaoxia Wu, Yiping Wang, Kuan Wang, Zheng Zhan, Olatunji Ruwase, Yelong Shen, Xin Eric Wang

preprint 2024

Arxiv

BibTeX

@misc{he2025mojitomotiontrajectoryintensity,
      title={Mojito: Motion Trajectory and Intensity Control for Video Generation}, 
      author={Xuehai He and Shuohang Wang and Jianwei Yang and Xiaoxia Wu and Yiping Wang and Kuan Wang and Zheng Zhan and Olatunji Ruwase and Yelong Shen and Xin Eric Wang},
      year={2025},
      eprint={2412.08948},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2412.08948}, 
}

Infer Human's Intentions Before Following Natural Language Instructions

Yanming Wan, Yue Wu, Yiping Wang, Jiayuan Mao, Natasha Jaques

AAAI 2025

Arxiv Code X Website

BibTeX

@misc{wan2024inferhumansintentionsfollowing,
      title={Infer Human's Intentions Before Following Natural Language Instructions}, 
      author={Yanming Wan and Yue Wu and Yiping Wang and Jiayuan Mao and Natasha Jaques},
      year={2024},
      eprint={2409.18073},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2409.18073}, 
}

CLIPLoss and Norm-Based Data Selection Methods for Multimodal Contrastive Learning

Yiping Wang*, Yifang Chen*, Wendan Yan, Alex Fang, Wenjing Zhou, Kevin Jamieson, Simon Shaolei Du

We design simple but efficient data selection methods for CLIP pretraining, and get new SOTA in DataComp benchmark.

NeurIPS 2024 (Spotlight)

Arxiv Code DataComp Poster X Previous Versions

BibTeX

@misc{wang2024cliplossnormbaseddataselection,
      title={CLIPLoss and Norm-Based Data Selection Methods for Multimodal Contrastive Learning}, 
      author={Yiping Wang and Yifang Chen and Wendan Yan and Alex Fang and Wenjing Zhou and Kevin Jamieson and Simon Shaolei Du},
      year={2024},
      eprint={2405.19547},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2405.19547}, 
}

JoMA: Demystifying Multilayer Transformers via JOint Dynamics of MLP and Attention

Yuandong Tian, Yiping Wang, Zhenyu Zhang, Beidi Chen, Simon Shaolei Du

We analyze the training dynamics of multilayer transformer, characterizing the role of self-attention and MLP nonlinearity.

ICLR 2024

Arxiv X

BibTeX

@misc{tian2024jomademystifyingmultilayertransformers,
      title={JoMA: Demystifying Multilayer Transformers via JOint Dynamics of MLP and Attention}, 
      author={Yuandong Tian and Yiping Wang and Zhenyu Zhang and Beidi Chen and Simon Du},
      year={2024},
      eprint={2310.00535},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2310.00535}, 
}

Scan and Snap: Understanding Training Dynamics and Token Composition in 1-layer Transformer

Yuandong Tian, Yiping Wang, Beidi Chen, Simon Shaolei Du

We analyze the 1-layer transformer with next token prediction loss, and rigorously prove its training process.

NeurIPS 2023 (Oral presentation @ ICML2023-HiDL)

Arxiv Poster X

BibTeX

@misc{tian2023scansnapunderstandingtraining,
      title={Scan and Snap: Understanding Training Dynamics and Token Composition in 1-layer Transformer}, 
      author={Yuandong Tian and Yiping Wang and Beidi Chen and Simon Du},
      year={2023},
      eprint={2305.16380},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2305.16380}, 
}

Improved Active Multi-Task Representation Learning via Lasso

Yiping Wang, Yifang Chen, Kevin Jamieson, Simon S. Du

ICML 2023

Arxiv

BibTeX

@misc{wang2023improvedactivemultitaskrepresentation,
      title={Improved Active Multi-Task Representation Learning via Lasso}, 
      author={Yiping Wang and Yifang Chen and Kevin Jamieson and Simon S. Du},
      year={2023},
      eprint={2306.02556},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2306.02556}, 
}

C-Mixup: Improving Generalization in Regression

Huaxiu Yao*, Yiping Wang*, Linjun Zhang, James Zou, Chelsea Finn

NeurIPS 2022

Arxiv Code

BibTeX

@misc{yao2022cmixupimprovinggeneralizationregression,
      title={C-Mixup: Improving Generalization in Regression}, 
      author={Huaxiu Yao and Yiping Wang and Linjun Zhang and James Zou and Chelsea Finn},
      year={2022},
      eprint={2210.05775},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2210.05775}, 
}