Yuan Lan

About Me

Yuan Lan 蓝缘 is currently a Researcher at Huawei Theory Lab in Hong Kong. Prior to joining Huawei in March 2023, she completed her Ph.D. in Mathematics at the Hong Kong University of Science and Technology (HKUST) under the guidance of Prof. Yang Xiang. She received her B.S. in Mathematics from Wuhan University in 2018.

Dr. Lan’s work lies at the intersection of mathematics, AI, and computer vision, with a particular interest in developing efficient, training-free algorithms for image generation/editing and physics-inspired computational methods.

Research Interests

• Generative Diffusion Models: Developing training-free methods for generative diffusion models, enabling universal image editing, inpainting, and conditional sampling across diverse model architectures.
• Image Processing: Leveraging physics-inspired approaches to solve challenges in image compression and segmentation.

Blog Posts

1. Rethinking the Diffusion Model from a Langevin Perspective
   Candi Zheng, Yuan Lan
   ICLR 2026 Blogpost Track
   📄 Blog📋 BibTeX

   @inproceedings{zheng2026rethinking,
     title={Rethinking the Diffusion Model from a Langevin Perspective},
     author={Candi Zheng and Yuan Lan},
     booktitle={The Blogpost Track at ICLR 2026},
     year={2026},
     url={https://openreview.net/forum?id=MLAN3xr4S7}
   }

Publications

1. Lanpaint: Training-Free Diffusion Inpainting with Exact and Fast Conditional Inference
   Candi Zheng*^∧, Yuan Lan^∧, Yang Wang
   Transactions on Machine Learning Research (TMLR), 2025
   📄 TMLR🐙 Code📋 BibTeX

   @article{zheng2025lanpaint,
     title={LanPaint: Training-Free Diffusion Inpainting with Asymptotically Exact and Fast Conditional Sampling},
     author={Candi Zheng and Yuan Lan and Yang Wang},
     journal={Transactions on Machine Learning Research},
     issn={2835-8856},
     year={2025},
     url={https://openreview.net/forum?id=JPC8JyOUSW}
   }

2. Characteristic Guidance: Non-linear Correction for Diffusion Model at Large Guidance Scale
   Candi Zheng*^∧, Yuan Lan^∧
   Proceedings of the 41st International Conference on Machine Learning (ICML), PMLR 235:61386-61412, 2024
   📄 ICML🐙 Code📋 BibTeX

   @inproceedings{zheng2024characteristic,
     title={Characteristic Guidance: Non-linear Correction for Diffusion Model at Large Guidance Scale},
     author={Candi Zheng and Yuan Lan},
     booktitle={Forty-first International Conference on Machine Learning},
     year={2024},
     url={https://openreview.net/forum?id=eOtjMYdGLt}
   }

3. ElasticLaneNet: A Geometry-Flexible Approach for Lane Detection
   Yaxin Feng, Yuan Lan, Luchan Zhang*, Yang Xiang*
   IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2025
   📄 WACV📋 BibTeX

   @inproceedings{10943462,
     author={Feng, Yaxin and Lan, Yuan and Zhang, Luchan and Xiang, Yang},
     booktitle={2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
     title={ElasticLaneNet: An Efficient Geometry-Flexible Lane Detection Framework},
     year={2025},
     pages={8744-8753},
     doi={10.1109/WACV61041.2025.00847}
   }

4. GOLLIC: Learning Global Context beyond Patches for Lossless High-Resolution Image Compression
   Yuan Lan, Liang Qin, Zhaoyi Sun, Yang Xiang*, Jie Sun*
   arXiv:2210.03301, 2022, Submitted
   📄 arXiv📋 BibTeX

   @misc{lan2022golliclearningglobalcontext,
     title={GOLLIC: Learning Global Context beyond Patches for Lossless High-Resolution Image Compression},
     author={Yuan Lan and Liang Qin and Zhaoyi Sun and Yang Xiang and Jie Sun},
     year={2022},
     eprint={2210.03301},
     archivePrefix={arXiv},
     primaryClass={eess.IV},
     url={https://arxiv.org/abs/2210.03301}
   }

5. DOSnet as a Non-Black-Box PDE Solver: When Deep Learning Meets Operator Splitting
   Yuan Lan, Zhen Li*, Jie Sun*, Yang Xiang*
   Journal of Computational Physics, 491, 112343, 2023
   📄 Journal📋 BibTeX

   @article{LAN2023112343,
     title={DOSnet as a non-black-box PDE solver: When deep learning meets operator splitting},
     journal={Journal of Computational Physics},
     volume={491},
     pages={112343},
     year={2023},
     issn={0021-9991},
     doi={https://doi.org/10.1016/j.jcp.2023.112343},
     url={https://www.sciencedirect.com/science/article/pii/S0021999123004382},
     author={Yuan Lan and Zhen Li and Jie Sun and Yang Xiang}
   }

6. Feature Flow Regularization: Improving Structured Sparsity in Deep Neural Networks
   Yue Wu, Yuan Lan, Luchan Zhang*, Yang Xiang*
   Neural Networks, 161, 598-613, 2023
   📄 Journal📋 BibTeX

   @article{WU2023598,
     title={Feature flow regularization: Improving structured sparsity in deep neural networks},
     journal={Neural Networks},
     volume={161},
     pages={598-613},
     year={2023},
     issn={0893-6080},
     doi={https://doi.org/10.1016/j.neunet.2023.02.013},
     url={https://www.sciencedirect.com/science/article/pii/S089360802300076X},
     author={Yue Wu and Yuan Lan and Luchan Zhang and Yang Xiang}
   }

7. An Elastic Interaction Based Loss Function for Medical Image Segmentation
   Yuan Lan, Yang Xiang, Luchan Zhang*
   International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI). Springer, Cham, 2020
   📄 MICCAI🐙 Code📋 BibTeX

   @InProceedings{10.1007/978-3-030-59722-1_73,
     author={Lan, Yuan and Xiang, Yang and Zhang, Luchan},
     title={An Elastic Interaction-Based Loss Function for Medical Image Segmentation},
     booktitle={Medical Image Computing and Computer Assisted Intervention -- MICCAI 2020},
     year={2020},
     publisher={Springer International Publishing},
     address={Cham},
     pages={755--764},
     isbn={978-3-030-59722-1}
   }

Note: *Corresponding authors, ^∧Equal contribution

Awards and Scholarships

• Huawei APRI Starlight Award, Huawei (2024)
• Huawei President’s Award of APRI 2024, Huawei (2024)
• Huawei Innovation Pioneer, First Prize, Huawei (2024H1)
• HKUST RedBird Academic Excellence Awards, HKUST (2021-2022)
• Huawei PHD Fellowship, Huawei Hong Kong (2019-2022)
• Huawei Innovation Pioneer, Second Prize, Huawei (2020H2)
• Third Class Din-Yu Hsieh Best Teaching Assistant Award, HKUST (2019-2020)
• Postgraduate Scholarship (PGS), HKUST (2018-2019)

Student Entrepreneurship

🚀 Founder, LuoBiBan (珞毕办), Wuhan University, 2016

Founded and launched LuoBiBan, a photographic platform connecting professional photographers with graduates for commemorative photography services. The platform operated as a marketplace similar to Uber, facilitating connections between photographers and students while charging agency fees. Successfully developed and sold exclusive WHU autograph albums. 📄 Featured in WHU News

Professional Activities

Talks:

• ICIAM 2023: DOSnet as a Non-Black-Box PDE Solver: When Deep Learning Meets Operator Splitting
  Tokyo, Japan, Aug 2023
• MICCAI Oral: An Elastic Interaction Based Loss Function For Medical Image Segmentation
  Lima, Peru (online), Oct 2020
• ICIAM 2019: Numerical Method for the Grain Boundary Dynamics Models with Long-range Elastic Interaction
  Valencia, Spain, Aug 2019

Teaching:

• TA Math 2011: Introduction to Multivariable Calculus (Spring 2020)
• TA Math 1013: Calculus IB (Fall 2019)
• TA Math 1014: Calculus II (Spring 2019)

Contact

• Email: [email protected]

Powered by Jekyll and Minimal Light theme.