5 papers
Retrieval Heads Meet Vision: Uncovering How VLMs Locate and Extract Visual Information
Chanho Park, Daehyeon Choi, Jihyun Lee +1
Vision-language models (VLMs) can locate an image region referred to by a text prompt and route the corresponding visual evidence to the output, yet the internal mechanism behind t…
DiffusionRollout: Uncertainty-Aware Rollout Planning in Long-Horizon PDE Solving
Seungwoo Yoo, Juil Koo, Daehyeon Choi +1
We propose DiffusionRollout, a novel selective rollout planning strategy for autoregressive diffusion models, aimed at mitigating error accumulation in long-horizon predictions of…
Toward Ambulatory Vision: Learning Visually-Grounded Active View Selection
Juil Koo, Daehyeon Choi, Sangwoo Youn +2
Vision Language Models (VLMs) excel at visual question answering (VQA) but remain limited to snapshot vision, reasoning from static images. In contrast, embodied agents require amb…
ORIGEN: Zero-Shot 3D Orientation Grounding in Text-to-Image Generation
Yunhong Min, Daehyeon Choi, Kyeongmin Yeo +2
We introduce ORIGEN, the first zero-shot method for 3D orientation grounding in text-to-image generation across multiple objects and diverse categories. While previous work on spat…
InstantDrag: Improving Interactivity in Drag-based Image Editing
Joonghyuk Shin, Daehyeon Choi, Jaesik Park
Drag-based image editing has recently gained popularity for its interactivity and precision. However, despite the ability of text-to-image models to generate samples within a secon…