Reproduce “HEIR: Harness Egocentric Intent for Human-Robot Interactions” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yichi Zhang, Bo Wang, Zhenghan Zhu, Yutong Han, Jiayin Zou, Gao Kejun, Chang Liu, Shaoxuan Xie and 6 more
HEIR records human gaze and speech alongside robot views and actions, linking them to requests and navigation or manipulation subtasks. The dataset supports research on long-horizon human–robot interaction. In an offline ablation, Gemma-4-31B reached 77.69% Location Accuracy with measured gaze, versus 71.39% with a center-prior proxy. Whole-request success was 18.52% across 27 episodes, compared with 81.25% navigation-subtask success across 32 subtasks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.