Dr. Wuyang Li

I will be a postdoctoral fellow at the AIM Group of The Chinese University of Hong Kong (CUHK), supervised by Prof. Yixuan Yuan in 2024. Prior to this, I completed my PhD degree (2020-2023) at the City University of Hong Kong (CityU) with early graduation and my Bachelor's degree (2016-2020) at Tianjin University. I am broadly interested in computer vision, machine learning, and graph-based learning. My PhD research is in line with graph-based learning for object detection. Currently, my primary focus revolves around large vision/language/multi-modal models, including both technical research and practical applications.

I am also looking for overseas opportunities in the U.S. in 2024.

I have established an startup team to create a fantastic educational product with AIGC. The related technologies consist of language/vision generation. We have started the project and successfully applied for the seed funding with hundreds of thousands of HK $. If you are interested in our startup team, please contact me by email 📧!!

Email  /  CV  /  Google Scholar  /  Github

• [2023/12] I pass my PhD defense (early graduation)!!
• [2023/08] 1 paper is selected as ORAL in ICCV.
• [2023/07] 2 papers are accepted by ICCV 2023.
• [2023/06] Successfully apply the seed funding for our entrepreneurial team!
• [2023/05] 1 paper is accepted by MICCAI 2023.
• [2023/04] 1 paper is accepted by TNNLS 2023.
• [2023/03] 1 paper is accepted by TMI 2023.
• [2023/02] 1 paper is accepted by CVPR 2023.
• [2023/01] 1 paper is accepted by TPAMI 2023.
• [2022/10] 1 paper is accepted by TMM 2022.
• [2022/06] SIGMA appears on CVPR Best Paper Finalist [33/8161]!
• [2022/05] 1 paper is accepted by MICCAI 2022 (Early Accept, ORAL).
• [2022/03] 2 papers are accepted by CVPR 2022 (one ORAL ).
• [2021/10] 1 paper is accepted by AAAI 2022 ( ORAL ).
• [2021/07] I passed Ph.D. Qualify Examination.

Key Words: Open Set/ Domain Adaptive Object Detection

Summary: We formulated a real-world friendly setting, called Adaptive Open-set Object Detection (AOOD), which considers both novel scenes and novel classes. Then, we addressed this by selecting subgraphs (motifs) among object queries for a high-order optimization.

Key Words: Multimodal Pretraining

Summary: We observed the failure in Masked Image Modeling (MIM) when the scale of the informative foreground is limited. To address this, we proposed Masked Relation Modeling (MRM), which masks the cross-token relations for reliable pretraining.

Key Words: Open Set Domain Adaptation, Causal Theory

Summary: We decoupled images into base-class and novel-class regions and addreseed the biased learning in the source domain and biased cross-domain transfer with causal theory.

Key Words: Domain Adaptive Object Detection, Graph-based Learning

Summary: We proposed a SemantIc-complete Graph MAtching (SIGMA) framework to address the cross-domain semantic-mismatching with fine-grained domain adaptation.

Key Words: Domain adaptive Object Detection, Noisy Label

Summary: We explored the robust adaptive object detection under noisy annotations and proposed a Noise Latent Transferability Exploration (NLTE) framework to address the issue.

Key Words: Domain adaptive Object Detection, Graph-based Learning

Summary: We proposed a Semantic Conditioned AdaptatioN (SCAN) framework to address the sub-optimal categorical alignment for domain adaptive object detection.

Key Words: Gyral Hinges Alignment, Hypergraph Matching

Summary: We proposed a multi-scale hypergraph matching framework to achieve a robust alignment between the gyral hinges from two brain MRIs.

Key Words: Federated Learning, Noisy Label, Causal Intervention

Summary: We proposed to study and explored the recognition bias in federated object detection under noisy annotations , and then addressed the bias with a novel causality-driven framework.

Key Words: Domain Adaptive Object Detection, Hypergraph-based Learning

Summary: We improved the original graph-matching framework (SIGMA) with the high-order hypergraph space, which addresses the cross-domain semantic-misalignment with fine-grained domain adaptation.

Key Words: Domain Adaptive Object Detection, Conditional Kernels

Summary: We proposed an enhanced semantic-condtioned framework to address the sub-optimal categorical alignment for cross-doamin detection.

Key Words: Generic Object Detection, Graph-based Learning

Summary: We proposed a Heterogeneous Task Decoupling (HTD) framework to disentangle the sibling head with Graph Convolutional Network and Convolutional Network of two-stage detection pipeline .

Key Words: Point-cloud Segmentation, Graph-based Learning

Summary: We used the graph to model point clouds and addressed the ambiguous boundary issue via graph-based learning.

Key Words: Source-free Domain Adaptive Object Detection, Causal Theory

Summary: We leveraged causal theory to explore the bias problem in SFDA, and addressed the sample bias, feature bias, and prediction bias.

• IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
• International Journal of Computer Visio (IJCV)
• IEEE Transactions on Neural Networks and Learning Systems (TNNLS)
• IEEE Transactions on Image Processing (TIP)
• IEEE Transactions on Automation Science and Engineering (TASE)
• Pattern Recognition (PR)
• IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2023, 2024
• IEEE International Conference on Computer Vision (ICCV), 2023
• AAAI Conference on Artificial Intelligence (AAAI), 2023, 2024

• [2023] Outstanding Academic Performance Award (OAPA), CityU
• [2023] Research Tuition Scholarship (RTS), CityU
• [2022] Outstanding Academic Performance Award (OAPA), CityU
• [2022] Research Tuition Scholarship (RTS), CityU
• [2018] National Scholarship (Top 2% student)
• [2017] National Scholarship (Top 2% student)
• [2017] Tianjin Mathematical Competition (Second Prize)
• [2017-2020] Outstanding Student Scholarship (Top 10% student)

Sep. 2016 - Jun. 2020： Bachelor's degree of Communication Engineering.

GPA: 3.83/4.00, 91.3/100, Ranking 6/120

Sep. 2019 - Jun. 2020: Exchanging program of Electrical and Computer Engineering.

Supervisors: Prof. Zhiying Zhou

Complete the project: Towards Webpage-based Augmentation Reality (AR)

Sep. 2020 - present: Ph.D Study of Electrical Engineering.

Supervisors: Prof. Yixuan Yuan

Jun. 2017 - Jun. 2018: Freshman leader for Class 2, communication engineering

I was fortunate to be one of eight freshman leaders selected through the departmet.

Sep. 2018 - Jun. 2020: Chairman of the publicity department.

I was fortunate to be selected as the publicity department chairman of the student union in Electrical and Information Engineering Department, Tianjin University.

• Painting and Desinging: I used to do sketch training with art candidates and have a certain level of graphic design foundation. I have a strong interest in user needs analysis and product design.
• I am looking for the opportunity to establish a start-up team and create some awesome high-tech products.