Associate Professor, Korea Advanced Institute of Science and Technology
3 papers at NeurIPS 2025
VIPAMIN is a lightweight visual prompt initialization method that enhances self-supervised model adaptation by aligning prompts with semantic regions and expanding representational diversity.
We propose SNAP, a novel low-latency Test-Time Adaptation framework that enables efficient model adaptation on edge devices by using sparse updates, significantly reducing computation while maintaining accuracy.
We propose CovMatch, a scalable dataset distillation method for image-text contrastive learning that aligns cross-covariance and feature distributions between real and synthetic data.
