This paper presents a framework for training an agent to actively request help in object-goal navigation tasks, with feedback indicating the location of the target object in its field of view. To make the agent more robust in scenarios where a teacher may not always be available, the proposed training curriculum includes a mix of episodes with and without feedback. The results show that this approach improves the agent's performance, even in the absence of feedback.
@article{arxiv.2303.15453,
title = {Robustness of Utilizing Feedback in Embodied Visual Navigation},
author = {Jenny Zhang and Samson Yu and Jiafei Duan and Cheston Tan},
journal= {arXiv preprint arXiv:2303.15453},
year = {2023}
}
Comments
Accepted at the ICRA Workshop for Communicating Robot Learning across Human-Robot Interaction