Yong-Min Shin
jordan3414@yonsei.ac.kr
Hello! I am a Ph.D candidate in MIDaS Lab led by Prof. Won-Yong Shin, Dept. Computer Science & Engineering in Yonsei University, South Korea. Before joining MIDaS, I graduated from Yonsei University in Feb. 2019 as a physics major.
My primary interest is to find elegant and simple solutions to interesting and important problems on graph data and graph neural networks (GNNs). More specifically, my main research topics are currently focused on…
- Explainabiliy in graph neural networks
- Efficient graph learning
- And possibly exploiting explanations to make graph learning more efficient!
I also have research experiences in…
- Time-series with INR
- Graph representation learning
- Explainable healthcare
- Recommendation systems
…which are published in conferences like AAAI, IJCAI, SIGIR, LoG, and journals like PAMI, PlosOne, and so on.
Recently, I am becoming increasingly interested in…
- Mechanistic Interpretability: I personally think that this is the future of XAI research. I just started to pick up some papers on this topic and I am trying to understand the concept of mechanistic interpretability. As a means to learn more about this topic, I am currently translating several important articles on mechanistic interpretability into Korean (Coming soon!).
- Causal Inference: With the help of Brady Neal’s Introduction to Causal Inference, I am trying to understand the concept of causality to understand how it can impact the explainability of machine learning models.
I am always open to new ideas and collaborations, so feel free to reach out to me!
Here is my formal CV :)
And I also have a personal blog. Check it out!
News
Jun 5, 2024 | The paper “On the Feasibility of Fidelity\(^{-}\) for Graph Pruning” was accepted in IJCAI 2024 Workshop on Explainable Artificial Intelligence (XAI) where I am the first author. Thanks to Prof. Shin for the supervision and to the reviewers! The paper is available here. |
---|---|
Mar 26, 2024 | The paper “Turbo-CF: Matrix Decomposition-Free Graph Filtering for Fast Recommendation” was accepted in SIGIR 2024 (Short track), where I participated as the second author by participating in the experiments (e.g., runtime measurements, baselines) and writing. Thanks to all the co-authors and the reviewers! The paper is available here. |
Mar 15, 2024 | The paper “PAGE: Prototype-Based Model-Level Explanations for Graph Neural Networks” was accepted in Transactions on Pattern Analysis and Machine Intelligence (TPAMI), where I am the first author. Thanks to all the co-authors and the reviewers! The paper is available here. |
Jan 15, 2024 | I will start my month-long journey to the University of New South Wales (UNSW) in Sydney, Australia, as a visiting researcher. I will be working with Prof. Xin Cao and his group on explanation methods for graph neural networks. I am excited to meet new people and learn new things! (Also, check out my blog post for the photos during the stay) |
Publications
Latest posts
Nov 11, 2022 | Comparing the Efficiency of ChebConv, GCN, and SGC |
---|---|
Oct 30, 2022 | Understanding ChebConv with the NEIGHBORSMATCH problem |
Oct 19, 2022 | Dissecting Cora in the Spectral Domain |