Prof. Jangho Kim - Kookmin University AI·SW Graduate School

Research Overview연구 개요

MLPR Lab researches efficient AI algorithms and models using Machine Learning and Pattern Recognition. Focus areas include model compression via quantization, pruning, and knowledge distillation, as well as machine unlearning and continual learning. The goal is 'Ubiquitous AI' - making AI accessible to everyone, everywhere.

MLPR 연구실은 머신러닝과 패턴인식을 활용하여 효율적인 AI 알고리즘과 모델을 연구합니다. 양자화, 프루닝, 지식 증류를 통한 모델 압축과 머신 언러닝, 지속 학습 등을 주요 연구 분야로 다루고 있습니다.

Research Areas연구 분야

Model Compression
Quantization
Knowledge Distillation
Machine Unlearning
Pruning
Efficient AI

모델 압축
양자화
지식 증류
머신 언러닝
프루닝
효율적 AI

Major Achievements주요 연구 성과

Nonlinear Color Transfer via Learnable Bezier Flows (CVPR 2026)
Feature-map-level Online Adversarial Knowledge Distillation (ICML 2020)
Paraphrasing Complex Network: Factor Transfer (NeurIPS 2018)

Honors & Awards수상

2nd Place, Machine Unlearning Challenge (NeurIPS 2023)
1st Place, DCASE Challenge Task 1-a
3rd Place, MicroNet Challenge (NeurIPS 2019)

Recent Publications주요 논문

Hae Eun Lee, Hyungi Lee*, Jangho Kim*, "PriSM: Prior-guided Shared-basis Mixture Personalization for LLMs under Sparse User Histories," Conference on Neural Information Processing Systems (NeurIPS 2026), 2026
Hyungi Lee, Jangho Kim, "Improving Neural Processes in the Low-Data Regime via Context-Subset Training and Self-Distillation," Conference on Neural Information Processing Systems (NeurIPS 2026), 2026
H. Cho, S. An, H. Park, and J. Kim, "SharedKD: Gradient-Guided Dynamic Student Discovery inside a Single 3D Object Detector," DAC, 2026
D. Lee, J. Shin, J. Bae, H. Cho, and J. Kim, "Weight Initialization based on Gradient Similarity for Versatile Machine Unlearning," PETS, 2026
J. Lee, S. Jo, J.-H. Park, Y. Ryou, J. Yang, and J. Kim, "Nonlinear Color Transfer via Learnable Bezier Flows," CVPR, 2026
H. Lee, H. Jo, J. Chung, and J. Kim, "SQuaT: Self-Supervised Knowledge Distillation via Student-Aware Quantized Teacher Features," AISTATS, 2026 원문 ↗Paper ↗
J. Kim, J. Shin, S. An, and J. Kim, "Exploring Diverse Sparse Network Structures via Dynamic Pruning with Weight Alignment," CIKM, 2025
J. Shin, J. Kim, and J. Kim, "Entropy-Guided Meta-Initialization regularization for few-shot text classification," Knowledge-Based Systems, 2025
S. An, J. Kim, K. Lee, J. Huh, C. Kwak, Y. Lee, M. Jin, and J. Kim, "Sparse Structure Exploration and Re-optimization for Vision Transformer," UAI, 2025
J. Back, N. Ahn, and J. Kim, "Magnitude attention-based dynamic pruning," Expert Systems with Applications, 2025
J. Kim, J. Lee, D. Han, and N. Kwak, "Layerwise-priority-based gradient adjustment for few-shot learning," Expert Systems with Applications, 2025
S. An, J. Shin, and J. Kim, "Quantization-aware training with Dynamic and Static Pruning," IEEE Access, 2025
J. Shin, S. Han, and J. Kim, "Cooperative Meta-Learning with Gradient Augmentation," UAI, 2024
J. Kim, J. Yoo, Y. Song, K. Yoo, and N. Kwak, "Finding Efficient Pruned Network via Refined Gradients for Pruned Weights," ACM MM, 2023
J. Jang, K. Yoo, S. Kim, C. Kong, J. Kim, and N. Kwak, "Self-Distilled Self-Supervised Representation Learning," WACV, 2023
J. Kim, "Quantization Robust Pruning With Knowledge Distillation," IEEE Access, 2023
J. Kim, J.-T. Lee, S. Chang, and N. Kwak, "Variational On-the-Fly Personalization," ICML, 2022
K. Yoo, J. Kim, J. Jang, and N. Kwak, "Detection of Word Adversarial Examples in Text Classification: Benchmark and Baseline via Robust Density Estimation," Findings of ACL, 2022
J. Kim, K. Yoo, and N. Kwak, "Position-based Scaled Gradient for Model Quantization and Sparse Training," NeurIPS, 2020
I. Chung, S. Park, J. Kim, and N. Kwak, "Feature-map-level Online Adversarial Knowledge Distillation," ICML, 2020
J. Kim, S. Park, and N. Kwak, "Paraphrasing Complex Network: Network Compression via Factor Transfer," NeurIPS, 2018