Changick Kim

Breaking Safety Alignment in Large Vision-Language Models via Benign-to-Harmful Optimization

Station2Radar: Query‑Conditioned Gaussian Splatting for Precipitation Field

BEEP3D: Box-Supervised End-to-End Pseudo-Mask Generation for 3D Instance Segmentation

Safety-Aligned Weights Are Not Enough: Refusal-Teacher-Guided Finetuning Enhances Safety and Downstream Performance under Harmful Finetuning Attacks

SELFI: Selective Fusion of Identity for Generalizable Deepfake Detection

Long-tailed Adversarial Training with Self-Distillation

Indirect Gradient Matching for Adversarial Robust Distillation

RITUAL: Random Image Transformations as a Universal Anti-hallucination Lever in LVLMs

Don't Miss the Forest for the Trees: Attentional Vision Calibration for Large Vision Language Models