Yu-Jhe Li (李宇哲)

• Apr 2026Joined Google as a Software Engineer working on multi-modal generative AI. 🎉
• Sep 2025One paper accepted to NeurIPS 2025 as Oral.
• Jul 2025One paper accepted to COLM 2025.
• Apr 2024Joined Adobe as a Research Scientist.
• Feb 2024One paper accepted to CVPR 2024.

My research focuses on multi-modal reasoning and visual generation, specifically:

• Leveraging multi-modal large language models (MLLMs) to enhance high-level reasoning, complex perception, and decision-making.
• Advancing generative AI and visual content generation—particularly through diffusion models—to synthesize high-fidelity images and videos for XR and autonomous systems.
• Developing data-efficient and adaptable methods that bridge the gap between cognitive reasoning and dynamic generation in real-world applications.

A curated list of representative works. See the full list on Google Scholar.

Understanding and Mitigating Numerical Sources of Nondeterminism in LLM Inference NeurIPS Oral

Jiayi Yuan, Hao Li, Xinheng Ding, Wenya Xie, Yu-Jhe Li, Wentian Zhao, Kun Wan, Jing Shi, Xia Hu, Zirui Liu

Conference on Neural Information Processing Systems (NeurIPS), 2025

arXiv

Efficient Self-Improvement in Multimodal Large Language Models: A Model-Level Judge-Free Approach

Shijian Deng, Wentian Zhao, Yu-Jhe Li, Kun Wan, Daniel Miranda, Ajinkya Kale, Yapeng Tian

Conference on Language Modeling (COLM), 2025

arXiv

Flexible Depth Completion for Sparse and Varying Point Densities

Jinhyung Park, Yu-Jhe Li, Kris Kitani

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024

Paper

Multi-Person 3D Pose Estimation from Multi-View Uncalibrated Depth Cameras

Yu-Jhe Li, Yan Xu, Rawal Khirodkar, Jinhyung Park, Kris Kitani

Preprint, arXiv, 2024

arXiv

Domain Adapted Visual Representation Learning for Machine Perception PhD Thesis

Yu-Jhe Li

Ph.D. Thesis, Carnegie Mellon University, 2023

Committee: Kris Kitani, Matthew O'Toole, Jun-Yan Zhu, Deva Ramanan

Thesis

Azimuth Super-Resolution for FMCW Radar in Autonomous Driving

Yu-Jhe Li, Shawn Hunt, Jinhyung Park, Matthew O'Toole, Kris Kitani

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023

Paper Dataset & Code

3D-CLFusion: Fast Text-to-3D Rendering with Contrastive Latent Diffusion

Yu-Jhe Li, Tao Xu, Ji Hou, Bichen Wu, Xiaoliang Dai, Albert Pumarola, Peizhao Zhang, Peter Vajda, Kris Kitani

Preprint, arXiv, 2023

arXiv

ST-MVDNet++: Improve Vehicle Detection with Lidar-Radar Geometrical Augmentation via Self-Training

Yu-Jhe Li, Matthew O'Toole, Kris Kitani

IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2023

Paper

3D-Aware Encoding for Style-based Neural Radiance Fields

Yu-Jhe Li, Tao Xu, Bichen Wu, Ningyuan Zheng, Xiaoliang Dai, Albert Pumarola, Peizhao Zhang, Peter Vajda, Kris Kitani

Preprint, arXiv, 2022

arXiv

Domain Adaptive Hand Keypoint and Pixel Localization in the Wild

Takehiko Ohkawa, Yu-Jhe Li, Qichen Fu, Ryosuke Furuta, Kris M. Kitani, Yoichi Sato

European Conference on Computer Vision (ECCV), 2022

arXiv

Cross-Domain Adaptive Teacher for Object Detection

Yu-Jhe Li, Xiaoliang Dai, Chih-Yao Ma, Yen-Cheng Liu, Kan Chen, Bichen Wu, Zijian He, Kris Kitani, Peter Vajda

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022

Project Paper Code

Modality-Agnostic Learning for Radar-Lidar Fusion in Vehicle Detection

Yu-Jhe Li, Jinhyung Park, Matthew O'Toole, Kris Kitani

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022

Paper

Visio-Temporal Attention for Multi-Camera Multi-Target Association

Yu-Jhe Li, Xinshuo Weng, Yan Xu, Kris Kitani

IEEE International Conference on Computer Vision (ICCV), 2021

CVF Paper

Wide-Baseline Multi-Camera Calibration using Person Re-Identification

Yan Xu, Yu-Jhe Li, Xinshuo Weng, Kris Kitani

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2021

CVF Paper

Learning Shape Representations for Person Re-Identification under Clothing Change

Yu-Jhe Li, Xinshuo Weng, Kris Kitani

Winter Conference on Applications of Computer Vision (WACV), 2021

arXiv CVF Paper

Transforming Multi-Concept Attention into Video Summarization

Yen-Ting Liu, Yu-Jhe Li, Yu-Chiang Frank Wang

Asian Conference on Computer Vision (ACCV), 2020

arXiv CVF Paper

Cross-Resolution Adversarial Dual Network for Person Re-Identification and Beyond

Yu-Jhe Li*, Yun-Chun Chen*, Yen-Yu Lin, Yu-Chiang Frank Wang  (* equal contribution)

Preprint, arXiv, 2020

arXiv

Cross-Dataset Person Re-Identification via Unsupervised Pose Disentanglement and Adaptation

Yu-Jhe Li, Ci-Siang Lin, Yan-Bo Lin, Yu-Chiang Frank Wang

IEEE International Conference on Computer Vision (ICCV), 2019

arXiv CVF Paper Code

Recover and Identify: A Generative Dual Model for Cross-Resolution Person Re-Identification

Yu-Jhe Li*, Yun-Chun Chen*, Yen-Yu Lin, Xiao-fei Du, Yu-Chiang Frank Wang  (* equal contribution)

IEEE International Conference on Computer Vision (ICCV), 2019

arXiv CVF Paper

Spot and Learn: A Maximum-Entropy Image Patch Sampler for Few-Shot Classification

Wen-Hsuan Chu, Yu-Jhe Li, Jing-Cheng Chang, Yu-Chiang Frank Wang

IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2019

CVF Paper

Learning Resolution-Invariant Deep Representations for Person Re-Identification Oral

Yu-Jhe Li*, Yun-Chun Chen*, Xiao-fei Du, Yu-Chiang Frank Wang  (* equal contribution)

AAAI Conference on Artificial Intelligence (AAAI), 2019

arXiv AAAI Paper

Deep Reinforcement Learning for Playing 2.5D Fighting Games

Yu-Jhe Li*, Hsin-Yu Chang*, Yu-Jing Lin, Po-Wei Wu, Yu-Chiang Frank Wang  (* equal contribution)

IEEE International Conference on Image Processing (ICIP), 2018

arXiv IEEE Paper Code

Adaptation and Re-Identification Network: An Unsupervised Deep Transfer Learning Approach to Person Re-Identification

Yu-Jhe Li, Fu-En Yang, Yen-Cheng Liu, Yu-Ying Yeh, Xiaofei Du, Yu-Chiang Frank Wang

IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2018

arXiv Paper Code