NoisyEQA: benchmarking Embodied Question Answering with imperfect queries from non-expert users
Robot Learning · 19.05.2026 · 1-25
DOI: 10.55092/rl20260014 ↗Kurz gefasst
Fachartikel in der Zeitschrift Robot Learning, online erschienen am 19.05.2026. Verfasst von Tao Wu und 5 weiteren Autor:innen.
Abstract (Originalsprache)
: Embodied Question Answering (EQA) enables robots to explore the environment and answer human questions, which is important for human-robot interaction and has been significantly enhanced by the recent advancement of Vision-Language Models (VLMs). However, EQA in real-world scenarios remains challenging, as human-posed questions often contain noise that can interfere with an embodied agent’s reasoning, bringing challenges for language beginners and non-expert users. To address this, we introduce a NoisyEQA benchmark designed to evaluate the ability of the robot to identify and correct noisy questions. NoisyEQA is inspired by three common types of noise observed in real-world applications: Memory Noise, Perception Noise, and Semantic Noise, generated through an automated dataset creation framework. Additionally, we propose a ‘Self-Correction’ prompting mechanism to enhance EQA against noise robustness and a novel evaluation metric to measure both noise detection capability and answer quality. Our comprehensive study reveals that current embodied agents often struggle to detect noise in questions, leading to responses that frequently contain incorrect information. Through our self-correct prompting mechanism, we can effectively improve the accuracy of agent answers.
Zitieren
APA
Wu, T., Zhou, C., Cao, H., Wong, Y. H., Gu, L., & Yang, J. (2026). NoisyEQA: benchmarking Embodied Question Answering with imperfect queries from non-expert users. Robot Learning, 1-25. https://doi.org/10.55092/rl20260014
Vancouver
Wu T, Zhou C, Cao H, Wong YH, Gu L, Yang J. NoisyEQA: benchmarking Embodied Question Answering with imperfect queries from non-expert users. Robot Learning. 2026:1-25. doi:10.55092/rl20260014
BibTeX
@article{Wu2026,
author = {Wu, Tao and Zhou, Chuhao and Cao, Haozhi and Wong, Yen Heng and Gu, Lin and Yang, Jianfei},
title = {NoisyEQA: benchmarking Embodied Question Answering with imperfect queries from non-expert users},
journal = {Robot Learning},
year = {2026},
pages = {1--25},
doi = {10.55092/rl20260014}
}
EUVIMED