Awesome Robotics Manipulation · full_paper

IVRA: Improving Visual-Token Relations for Robot Action Policy with Training-Free Hint-Based Guidance

作者：Jongwoo Park, Kanchana Ranasinghe, Jinhyeok Jang, Cristina Mata, Yoo Sung Jang, Michael S Ryoo · 单位：Stony Brook University · 会议/期刊：arXiv · 日期：2026-01-22 · 来源：High-Level Structured Planning / Multimodal Reasoning / Affordance Reasoning

任务规划多模态推理可供性遥操作感知
新标签打开 AlphaXiv新标签打开 Paper新标签打开 Project
IVRA: Improving Visual-Token Relations for Robot Action Policy with Training-Free Hint-Based Guidance figure
AlphaXiv 中文论文页面（可滚动查看）