Giyeong Oh

Latest

• Random Is Hard to Beat: Active Selection in Online DPO with Modern LLMs
• Diffusion-Driven Two-Stage Active Learning for Low-Budget Semantic Segmentation