I am a researcher at Google DeepMind where I build SOTA video generation model (CV). Currently, I am interested in few-step generation approachs to address the iterative nature of diffusion models.

Before that, I was at Adobe Research, where I worked on both diffusion step distillation and knowledge distillation. I contributed to the early development of Adobe's Firefly model family, including Firefly Image 1 and Firefly Image 2.

I got my Ph.D. in the Department of Electrical and Computer Engineering at Princeton University, fortunately supervised by Prof. Sun-Yuan Kung and Prof. David Wentzlaff. I received my BEng in Electronic and Computer Engineering from Hong Kong University of Science and Technology (HKUST).

Publication

• Stepwise Credit Assignment for GRPO on Flow-Matching Models
  Yash Savani, Branislav Kveton, Yuchen Liu, Yilin Wang, Jing Shi, Subhojyoti Mukherjee, Nikos Vlassis, Krishna Kumar Singh
  To appear in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2026).
  (arxiv)


• Rethinking Global Text Conditioning in Diffusion Transformers
  Nikita Starodubcev, Daniil Pakhomov, Zongze Wu, Ilya Drobyshevskiy, Yuchen Liu, Zhonghao Wang, Yuqian Zhou, Zhe Lin, Dmitry Baranchuk
  To appear in The Fourteenth International Conference on Learning Representations (ICLR 2026).
  (arxiv)


• X-Fusion: Introducing New Modality to Frozen Large Language Models
  Sicheng Mo, Thao Nguyen, Xun Huang, Siddharth Srinivasan Iyer, Yijun Li, Yuchen Liu, Abhishek Tandon, Eli Shechtman, Krishna Kumar Singh, Yong Jae Lee, Bolei Zhou, Yuheng Li
  Best Paper of IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshop of Transformer for Vision (CVPRW 2025).
  (paper) (webpage)


• ZipIR: Latent Pyramid Diffusion Transformer for High-Resolution Image Restoration
  Yongsheng Yu, Haitian Zheng, Zhifei Zhang, Jianming Zhang, Yuqian Zhou, Connelly Barnes, Yuchen Liu, Wei Xiong, Zhe Lin, Jiebo Luo
  arXiv preprint arXiv:2504.08591 (2025).
  (paper)


• Generating, Fast and Slow: Scalable Parallel Video Generation with Video Interface Networks
  Bhishma Dedhia, David Bourgin, Krishna Kumar Singh, Yuheng Li, Yan Kang, Zhan Xu, Niraj K Jha, Yuchen Liu
  International Conference on Computer Vision (ICCV 2025).
  (paper) (webpage)


• DOLLAR: Few-Step Video Generation via Distillation and Latent Reward Optimization
  Zihan Ding, Chi Jin, Difan Liu, Haitian Zheng, Krishna Kumar Singh, Qiang Zhang, Yan Kang, Zhe Lin, Yuchen Liu
  International Conference on Computer Vision (ICCV 2025).
  (paper) (webpage)


• Mixture of efficient diffusion experts through automatic interval and sub-network selection
  Alireza Ganjdanesh, Yan Kang, Yuchen Liu, Richard Zhang, Zhe Lin, Heng Huang
  European Conference on Computer Vision (ECCV 2024).
  (paper) (arxiv)


• Attention-Driven Training-Free Efficiency Enhancement of Diffusion Models
  Hongjie Wang, Difan Liu, Yan Kang, Yijun Li, Zhe Lin, Niraj Jha, Yuchen Liu
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2024).
  (paper) (arxiv) (webpage) (video)


• SNED: Superposition Network Architecture Search for Efficient Video Diffusion Model
  Zhengang Li, Yan Kang, Yuchen Liu, Difan Liu, Tobias Hinz, Feng Liu, Yanzhi Wang
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2024).
  (paper) (arxiv)


• Personalized Residuals for Concept-Driven Text-to-Image Generation
  Cusuh Ham, Matthew Fisher, James Hays, Nicholas Kolkin, Yuchen Liu, Richard Zhang, Tobias Hinz
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2024).
  (paper) (arxiv)


• Building Efficient Neural Prefetcher
  Yuchen Liu, Georgios Tziantzioulis, David Wentzlaff
  International Symposium on Memory Systems (MEMSYS 2023).
  (paper)


• 3D-FM GAN: Towards 3D-Controllable Face Manipulation
  Yuchen Liu, Zhixin Shu, Yijun Li, Zhe Lin, Richard Zhang, and S.Y. Kung
  European Conference on Computer Vision (ECCV 2022).
  (paper) (arxiv) (code) (webpage) (video) (poster)


• Evolving Transferable Neural Pruning Functions
  Yuchen Liu, S.Y. Kung, and David Wentzlaff
  Genetic and Evolutionary Computation Conference (GECCO 2022).
  (paper) (arxiv) (code)
  


• Class-Discriminative CNN Compression
  Yuchen Liu, David Wentzlaff, and S.Y. Kung
  International Conference on Pattern Recognition (ICPR 2022).
  (paper) (code)
  


• Content-Aware GAN Compression
  Yuchen Liu, Zhixin Shu, Yijun Li, Zhe Lin, Federico Perazzi, and S.Y. Kung
  IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR 2021).
  (paper) (arxiv) (webpage) (code)
  


• Rethinking Class-Discrimination Based CNN Channel Pruning
  Yuchen Liu, David Wentzlaff, and S.Y. Kung
  arXiv preprint arXiv:2004.14492 (2020).
  (paper)
  


• Methodical Design and Trimming of Deep Learning Networks: Enhancing External BP Learning with Internal Omnipresent-supervision Training Paradigm
  S.Y. Kung, Zejiang Hou, and Yuchen Liu
  IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP 2019).
  (paper)


• Extracting gait velocity and stride length from surrounding radio signals
  Chen-Yu Hsu, Yuchen Liu, Zachary Kabelac, Rumen Hristov, Dina Katabi, Christine Liu
  ACM Conference on Human Factors in Computing Systems (CHI 2017).
  (paper)

Tech Transfers

• Adobe Firefly Image 2, core researcher in model efficiency (link)
• Adobe Firefly Image 1 (link)

Service

• Conference Reviewer/Program Committee
  • CVPR - 2021, 2023, 2024, 2025
  • NeurIPS - 2021, 2022, 2023, 2024
  • ICML - 2022, 2023, 2024, 2025
  • ICLR - 2024, 2025
  • ECCV/ICCV - 2024, 2025
  • WACV - 2023, 2024
  • AAAI - 2023, 2024
• Journal Reviewer
  • ACM Computing Surveys

Interns and Collaborators

• Bhishma Dedhia, Princeton University
• Zihan Ding, Princeton University
• Hongwu Peng, University of Connecticut
• Dung Dinh, University of Syndney
• Sicheng Mo, UCLA
• Yongsheng Yu, Univeristy of Rochester
• Hongjie Wang, Princeton University
• Alireza Ganjdanesh, University of Maryland
• Zhengang Li, Northeastern University
• Cusuh Ham, Georgia Institute of Technology
• Jiaxin Xie, Hong Kong University of Science and Technology

Awards and Honors

• Best Paper Awards in CVPRW May 2025
  X-Fusion: Introducing New Modality to Frozen Large Language Models.
• HKUST Outstanding Undergraduate May 2017
  Awareded to top 3% of graduating undergraduate students
• The 14th National Challenge Cup, National Round, Third Prize October 2015
  Innovation competition joined by more than 2.5 million students from over 3,000 institutions
• Scholarship Scheme for Continuing Undergraduate Students 2013 - 2017
  Awarded to the top 5% of undergraduate students
• Dean's List 2014 - 2017
  Acknowledgement from HKUST's dean to students with excellent academic performance