About Me
Hi there, I'm Yu Wang, just got my Ph.D. degree from Visualization and Graphics Group, Utrecht University, under the supervision of Prof. Alexandru Telea. My research interests include high-dimensional data visualization, adversarial training (GAN), and machine learning classifier visualization.
Fun fact: I am an
FGA-certified gemologist
with expertise in gemstone identification.
Research
Inverse Projection and Decision Maps
- Projections:
- A projection reduces high-dimensional data to lower dimensions, mapping data from \( \mathbb{R}^n \) to \( \mathbb{R}^q \), often using \( q = 2 \) for visualization purposes (e.g., 2D scatterplots).
- Different projection techniques preserve data relationships with varying levels of computational speed and noise robustness.
- Inverse Projections:
- Inverse projections attempt to reverse the mapping of a projection, approximating the original high-dimensional data from its low-dimensional representation in \( \mathbb{R}^q \).
- They are generally more complex due to the non-injective nature of projections and are used in tasks like data imputation and shape morphing.
- Decision Maps:
- Decision maps visually represent the decision boundaries of machine learning models by coloring pixels in a 2D space, showing how models categorize data points.
- They are helpful in understanding model behavior and decision boundaries in explainable AI.
Selected Publications
-
Inverting Multidimensional Scaling Projections Using Data Point Multilateration
Daniela Blumberg, Yu Wang, Alexandru Telea, Daniel A. Keim, Frederik L. Dennig.
EuroVis workshop on visual analytics (EuroVA), 2024 -
Computing fast and accurate decision boundary maps
Cristian Grosu, Yu Wang, Alexandru Telea.
EuroVis workshop on visual analytics (EuroVA), 2024 -
Interpreting mineral deposit genesis classification with decision maps: a case study using pyrite trace elements
Yu Wang, Kun-Feng Qiu, Alexandru C. Telea, Zhao-Liang Hou, Tong Zhou, Yi-Wei Cai, Zheng-Jiang Ding, Hao-Cheng Yu, Jun Deng.
American Mineralogist, 2024 -
Seeing is Learning in High Dimensions: The Synergy Between Dimensionality Reduction and Machine Learning
Alexandru Telea, Alister Machado, Yu Wang.
SN Computer Science, 2024 -
Fundamental Limitations of Inverse Projections and Decision Maps (best student paper award)
Yu Wang, Alexandru Telea.
Proc. IVAPP, 2024 -
Quantitative and Qualitative Comparison of Decision-Map Techniques for Explaining Classification Models
Yu Wang, Alister Machado, Alexandru Telea.
Algorithms, 2023 -
Machine Learning Prediction of Quartz Forming-Environments
Yu Wang, KunFeng Qiu, Axel Müller, ZhaoLiang Hou, ZhiHai Zhu, HaoCheng Yu.
Journal of Geophysical Research: Solid Earth, 2021