About Me

I’m a Research Scientist @ SAIT, Samsung Electronics.

My research interest includes:

• Post-hoc analysis & applications of pre-traiend models (especially, Foundation Models).
• Generalization of deep learning via the lens of loss landscape.
• Bayesian deep learning (especially, Laplace Approximation).
• Explainability and interpretability of large-scale neural networks.

Education

• Ph.D. in Graduate School of AI, Korea Advanced Institute of Science and Technology (KAIST)
  Mar. 2019 - Feb. 2024

• M.S. in Industrial Engineering (Mathematical optimization), Seoul National University (SNU)
  Mar. 2017 - Feb. 2019

• B.S. in Industrial Engineering & Mathematical Science, Seoul National University (SNU)
  Mar. 2013 - Feb. 2017

Published Papers

• GEX: A flexible method for approximating influence via Geometric Ensemble
  Sung-Yub Kim, Kyungsu Kim, Eunho Yang
  NeurIPS, 2023

• RGE: A Repulsive Graph Rectification for Node Classification via Influence
  Jaeyun Song*, Sung-Yub Kim*, Eunho Yang (*: equal contribution)
  ICML, 2023

• Scale-invariant Bayesian Neural Networks with Connectivity Tangent Kernel
  Sung-Yub Kim, Sihwan Park, Kyungsu Kim, Eunho Yang
  ICLR, 2023 (Spotlighted)

• Generalized Tsallis Entropy Reinforcement Learning and Its Application to Soft Mobile Robots
  Kyungjae Lee, Sung-Yub Kim, Sungbin Lim, Sungjoon Choi, Mineui Hong, Jaein Kim, Yong-Lae Park, and Songhwai Oh
  Robotics: Science and Systems Foundation

Preprints

• TTD: Text-Tag Self-Distillation Enhancing Image-Text Alignment in CLIP to Alleviate Single Tag Bias
  Sanghyun Jo*, Soohyun Ryu*, Sungyub Kim, Eunho Yang, Kyungsu Kim (*: equal contribution)

• Post-Training Recovery from Injected Bias with Self-Influence
  Jaeyun Song* and Yeonsung Jung* and June Yong Yang* and Jin-Hwa Kim and Sung-Yub Kim and Eunho Yang (*: equal contribution)

• Bias Decay Matters: Improving Large Batch Optimization with Connectivity Sharpness
  Sung-Yub Kim, Sihwan Park, Yong-Deok Kim, Eunho Yang

• Reliable Estimation of Individual Treatment Effect with Causal Information Bottleneck
  Sung-Yub Kim, Yongsu Baek, Sung Ju Hwang, Eunho Yang

Projects

• Autonomous Intelligent Digital Companion Framework and Application
  Korea Electronics Technology Institute, 2019-2021

• A study on adaptive point and regional OOD detection techniques based on sub-task generation
  Samsung, 2020-2023

• A machine learning and statistical inference framework for explainable artificial intelligence
  Institute of Information & Communications Technology Planning & Evaluation (IITP), 2022-2023

• A Study on Optimization and Network Interpretation Method for Large-Scale Machine Learning
  National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT), 2022-2023