Website · CV · Scholar · LinkedIn
I am an AI researcher at SAIGE, working on industrial computer vision. I specialize in image anomaly detection and efficient training and inference, with a strong interest in incorporating language into vision systems for data-efficient and context-aware solutions. During my M.S. at KAIST, I studied the loss landscape of neural networks.
- Image anomaly detection
- Vision-language and language-assisted vision
- Efficient training and inference for vision systems
- Loss landscapes of neural networks
- ELDET: Early-Learning Distillation with Noisy Labels for Object Detection — NeurIPS 2025
- Language-Assisted Feature Transformation for Anomaly Detection — ICLR 2025
- A Simple Early Exiting Framework for Accelerated Sampling in Diffusion Models — ICML 2024
- Traversing Between Modes in Function Space for Fast Ensembling — ICML 2023
- Probabilistic Imputation for Time-series Classification with Missing Data — ICML 2023
- Martingale Posterior Neural Processes — ICLR 2023 Spotlight
- Scale Mixtures of Neural Network Gaussian Processes — ICLR 2022