|
I am a second-year (starting from 2023) Ph.D. student at University of California, Santa Cruz, supervised by Prof. Cihang Xie and Prof. Yuyin Zhou. Before joining UCSC, I was a researcher at Tencent. Previously, I got the Master degree at SUN YAT-SEN UNIVERSITY in 2021, supervised by Prof. Wei-Shi Zheng. Before that, I received my B.Eng. from Sichuan University in 2019. My research interests lie in deep learning and computer vision. I have published several works on image and video perception as well as multimodal learning. Currently, I focus on generative AI and foundation model training. Feel free to reach out to me at jinruiyang.ray@gmail.com for discussions or opportunities. I'm looking for research intern positions for Summer 2025.
CV / Google Scholar / Github |
|
|
|
|
Research Intern, 2024.6 - 2025.5 (Expected) Advisor: Qing Liu and Zhe Lin |
|
Researcher, 2021.7 - 2023.8 |
|
|
(*: Equal contribution) |
|
Jinrui Yang, Qing Liu, Yijun Li, Soo Ye Kim, Shilong Zhang, Daniil Pakhomov, Mengwei Ren, Jianming Zhang, Zhe Lin, Cihang Xie, Yuyin Zhou CVPR, 2025 paper/ page LayerDecomp outputs photorealistic clean backgrounds and high-quality transparent foregrounds with faithfully preserved visual effects. |
|
Jinrui Yang, Xianhang Li, Druv Pai, Yuyin Zhou, Yi Ma, Yaodong Yu, Cihang Xie NeurIPS, 2024 paper/ page/ code/ 新智元 We scaled the white-box transformer architecture, resulting in the CREATE-α model, which significantly improves upon the vanilla white-box transformer while preserving interpretability. This advancement nearly closes the gap between white-box transformers and ViTs. |
|
Chaoyou Fu, Peixian Chen, Yunhang Shen, Yulei Qin, Mengdan Zhang, Xu Lin, Jinrui Yang, Xiawu Zheng, Ke Li, Xing Sun, Yunsheng Wu, Rongrong Ji Preprint, 2023 paper/ page/ code/ 新智元 The paper introduces the first comprehensive Multimodal Large Language Model (MLLM) evaluation benchmark, MME. It measures both perception and cognition abilities across 14 subtasks, with 30 advanced MLLMs evaluated comprehensively on MME. |
|
Jinrui Yang, Jiawei Zhang, Fufu Yu, Xinyang Jiang, Mengdan Zhang, Xing Sun, Ying-Cong Chen, Wei-Shi Zheng ICCV, 2021 paper We propose a novel method to discretize pose information into visibility labels for body parts, ensuring robustness against sparse and noisy pose data in Occluded Person Re-ID. |
|
Jinrui Yang, Wei-Shi Zheng, Qize Yang, Ying-Cong Chen, Qi Tian CVPR, 2020 paper We propose a novel Spatial-Temporal Graph Convolutional Network (STGCN) to address the occlusion problem and the visual ambiguity issue caused by visually similar negative samples in video-based person Re-ID. |
|
Reviewer
|
|
|
Design and source code from Jon Barron's website |