| permalink | title | excerpt | author_profile | redirect_from | ||
|---|---|---|---|---|---|---|
/ |
true |
|
{% if site.google_scholar_stats_use_cdn %} {% assign gsDataBaseUrl = "https://cdn.jsdelivr.net/gh/" | append: site.repository | append: "@" %} {% else %} {% assign gsDataBaseUrl = "https://raw.githubusercontent.com/" | append: site.repository | append: "/" %} {% endif %} {% assign url = gsDataBaseUrl | append: "google-scholar-stats/gs_data_shieldsio.json" %}
My name is Wei Dai (戴威), I'm a 3th-year undergraduate student at the School of Communication Engineering, Xi'an University of Electronic Science and Technology, supervised by Licheng Jiao Google Scholar、Xinbo Gao Google Scholar、Xi Yang Google Scholar. I also worked with PhD student at Luojiashan Laboratory, Wuhan University Wenke Huang, PhD student at the School of Artificial Intelligence, Xi'an University of Electronic Science and Technology Yanbiao Ma.
🌟 I am actively applying for a PhD in the fall of 2026 , and applying for a direct PhD at a university summer camp in 2025!
🌟 I'm currently actively seeking for Ph.D., Research Assistant, or internship positions related to manifold Learning. I'm also open to any possible discussions or collaborate opportunities. If you're interested in working with me, don't hesitate to drop me an email!
"Any data is distributed in manifold space, and we use manifolds to explain objective physical laws"
I am passionate about using geometric perspectives to explain laws in the world of manifolds. Currently, my research interests mainly focus on the following aspects:
a) unfairness & Prejudice
b) Trustworthy Machine Learning
c) manifold learning
d) diffusion model
- 2024.12: 🚀🚀🚀 We will continue the work of 2025Tpami and solve the problem of misalignment between online enhanced semantic ratios and linear label ratios in manifold space.
- 2024.12: 🚀🚀🚀 We will continue the work of 2025CVPR and comprehensively analyze federated learning data augmentation from a geometric perspective in manifold space. In the following work, we will focus on class prototypes, Gaussian enhancement, and perceptual manifold geometric features.
- 2024.12: 🚀🚀🚀 We are enthusiastic about achieving the transition from pixel layer to image layer in segmentation tasks. In the future, we will carry out region instance prototype distribution recovery for long tail segmentation scenes from manifold space.
- 2024.12: 🚀🚀🚀 This is a major challenge!!! In our work on clip and dinov2, we discovered the non-uniformity of class prototypes, which will lead to a series of problems and hinder the development of large-scale models. In the future, we will focus on manifold space, try to explain the underlying mechanisms, and solve this bias.
- 2024.11: 🚀🚀🚀 Inspired by the alignment strategy of Xie Saining's team, we will further investigate the bias issue of diffusion models.
† Equal Contribution
- Geometric Knowledge-Guided Localized Global Distribution Alignment for Federated Learning
- Yanbiao Ma†, Wei Dai† (co-first), Wenke Huang, Jiayi Chen
- Conference on Computer Vision and Pattern Recognition (CCF-A CVPR), 2025
- Pursuing Better Decision Boundaries for Long-Tailed Object Detection via Category Information Amount
- Yanbiao Ma, Wei Dai, Jiayi Chen
- International Conference on Learning Representations (Tsinghua-A ICLR), 2025
- Guideline for Novel Fine‐Grained Sentiment Annotation and Data Curation: A Case Study
- Wei Dai, Wanqiu Kong, Tao Shang, Jianhong Feng, Jiaji Wu, Tan Qu
- CCF-C SCI-Q2 IF-2.84 ( journal 《Expert System》 ), 2025
- Trade-Offs Between Richness and Bias of Augmented Data in Long-Tailed Recognition
- Wei Dai, Yanbiao Ma, Jiayi Chen
- Under review, 2025
- Exploring Beyond Logits: Hierarchical Dynamic Labeling Based on Embeddings for Semi-Supervised Classification
- Jiayi Chen, Yanbiao ma, Wei Dai
- Under review, 2025
- Multi feature, modular artificial intelligence recognition method and system based on deep learning [Baidu Scholar] [Github]
- Wei Dai, Jiayi Chen, Fengliang Mi
- Under acceptance, 2024
- Multidimensional scoring system and method based on entropy weight method and statistical bias principle [Baidu Scholar]
- Jiayi Chen, Wei Dai, Fengliang Mi
- Under acceptance, 2024
- A personalized heart rate anomaly monitoring method based on LSTM [Baidu Scholar]
- Qiang Cui, Wei Dai, Jiayi Chen, Wenting Wei
- Under acceptance, 2024
- ✨✨✨ National level Leader of "Intelligent Sports Vision - Sports Training Feedback System Based on Deep Learning"
- ✨✨✨ National level Core member of "Unmanned Safety Operation Monitoring, Binding, and Emergency Command System"
- ✨✨✨ College Student Computer Innovation Competition National level third prize.
- ✨✨✨ mathematical contest in modeling provincial level second prize.
- ✨✨✨ Xi'an University of Electronic Science and Technology school level scholarship.
- Reviewer for ICLR'2025
- Reviewer for CVPR'2025
- 2022.09 - Now, Undergraduate, School of Communication Engineering, Xi'an University of Electronic Science and Technology, China (GPA:3.9/4.0, Weighted Average scorc:91/100)