PHIL framework: Operator early takeover corrects robot policies
What happened
Xinkai Jiang and other authors propose the PHIL framework, which lets operators preview a policy's predicted future actions via an XR interface before the robot executes them, take over to correct before bad actions occur, and pair rejected predictions with corrective actions for preference post-training. On four real manipulation tasks, with five correction rounds per condition, average success rate improved from 44.0% to 70.3% for diffusion policies and from 57.8% to 76.9% for X-VLA, with HG-DAgger as the comparison baseline. The authors state that policies trained with active corrections show fewer unnecessary recovery behaviors in successful episodes.
AI 根据报道生成 · 6 days ago更新
Reporting timeline
Follow the reporting to explore different perspectives.
- arXiv cs.ROPHIL: Proactive human-in-the-loop intervention via preference learning
The paper proposes the PHIL framework, which lets operators preview a policy's predicted future actions through an XR interface, take over to correct before undesirable actions occur, and pair rejected predictions with corrective actions for preference post-training. On four real manipulation tasks, with five correction episodes per condition, mean success rates rose from 44.0% to 70.3% for the diffusion policy and from 57.8% to 76.9% for X-VLA, compared with the HG-DAgger baseline; policies trained with proactive corrections showed fewer unnecessary recovery behaviors in successful episodes.
Activity for this event
- Current
- 3
- Peak
- 10Nov 5 14:00
- Past 24 hours
- -49%
The trend compares participants observed throughout the same period, so its range may be smaller than the current interest count. Move or tap to inspect each hour; use the arrow keys on a keyboard.