In-situ Value-aligned Human-Robot Interactions with Physical Constraints

TL;DR

This paper presents a framework integrating human preferences and physical constraints for robots using large language models, validated through a new household activity benchmark and experiments demonstrating effective task planning.

Contribution

It introduces In-Context Learning from Human Feedback (ICLHF) for balancing preferences and constraints, along with a household activity benchmark for evaluation.

Findings

ICLHF effectively generates task plans respecting preferences and constraints

The benchmark enables standardized evaluation of human-centered robot tasks

Experiments show improved task completion aligning with human feedback

Abstract

Equipped with Large Language Models (LLMs), human-centered robots are now capable of performing a wide range of tasks that were previously deemed challenging or unattainable. However, merely completing tasks is insufficient for cognitive robots, who should learn and apply human preferences to future scenarios. In this work, we propose a framework that combines human preferences with physical constraints, requiring robots to complete tasks while considering both. Firstly, we developed a benchmark of everyday household activities, which are often evaluated based on specific preferences. We then introduced In-Context Learning from Human Feedback (ICLHF), where human feedback comes from direct instructions and adjustments made intentionally or unintentionally in daily life. Extensive sets of experiments, testing the ICLHF to generate task plans and balance physical constraints with preferences, have demonstrated the efficiency of our approach. Project page: https://iclhf.github.io .