배준현 · Jun-Hyun Bae · Junhyun Bae

AI/ML PhD Candidate @ Kyungpook National University

Research Interests

My research focuses on understanding and improving the robustness and interpretability of deep learning models. My early work explored modular architectures for systematic generalization and associative memory, investigating how independent modules can learn recomposable representations for compositional reasoning. Building on this, I studied out-of-distribution (OOD) generalization, applying modular representations and causal inference frameworks to overcome spurious correlations and dataset biases in areas including medical imaging and debiasing. More recently, my interest has shifted to mechanistic interpretability — analyzing how semantic concepts are encoded within neural network components, such as cross-attention mechanisms in text-to-image diffusion models, enabling targeted interventions without retraining.

Education

Publications

Competitions

  • NeurIPS 2023 Machine Unlearning Challenge — 8th place (out of 1,188 teams), Google
  • AI Hackathon for Fashion Coordination (2020) — 3rd place, ETRI
  • AI Hackathon for Speech Recognition (2019) — 10th place, NAVER

Scholarships & Fellowships

  • CMU Visiting Scholar Program (2022) — Fully funded by Korean Government (IITP)
  • Full-Ride Scholarship (2019 – 2023) — Graduate, Academic Excellence, KNU
  • KNU+ 도전장학생 (2015 – 2019) — Merit Scholarship for Outstanding Entrants, Full tuition + stipend

Academic Service

Reviewer

  • ICML 2026 (Gold Reviewer)
  • ICML 2026 Mechanistic Interpretability Workshop
  • AAAI 2026

Contact