Where Did It Go Wrong? Capability-Oriented Failure Attribution for Vision-and-Language Navigation Agents

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Jianming, Wang, Yawen, Wang, Junjie, Xie, Xiaofei, Li, Shoubin, Wang, Qing, Xu, Fanjiang
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918470912835584
author Chen, Jianming
Wang, Yawen
Wang, Junjie
Xie, Xiaofei
Li, Shoubin
Wang, Qing
Xu, Fanjiang
author_facet Chen, Jianming
Wang, Yawen
Wang, Junjie
Xie, Xiaofei
Li, Shoubin
Wang, Qing
Xu, Fanjiang
contents Embodied agents in safety-critical applications such as Vision-Language Navigation (VLN) rely on multiple interdependent capabilities (e.g., perception, memory, planning, decision), making failures difficult to localize and attribute. Existing testing methods are largely system-level and provide limited insight into which capability deficiencies cause task failures. We propose a capability-oriented testing approach that enables failure detection and attribution by combining (1) adaptive test case generation via seed selection and mutation, (2) capability oracles for identifying capability-specific errors, and (3) a feedback mechanism that attributes failures to capabilities and guides further test generation. Experiments show that our method discovers more failure cases and more accurately pinpoints capability-level deficiencies than state-of-the-art baselines, providing more interpretable and actionable guidance for improving embodied agents.
format Preprint
id arxiv_https___arxiv_org_abs_2604_25161
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Where Did It Go Wrong? Capability-Oriented Failure Attribution for Vision-and-Language Navigation Agents
Chen, Jianming
Wang, Yawen
Wang, Junjie
Xie, Xiaofei
Li, Shoubin
Wang, Qing
Xu, Fanjiang
Multiagent Systems
Artificial Intelligence
Embodied agents in safety-critical applications such as Vision-Language Navigation (VLN) rely on multiple interdependent capabilities (e.g., perception, memory, planning, decision), making failures difficult to localize and attribute. Existing testing methods are largely system-level and provide limited insight into which capability deficiencies cause task failures. We propose a capability-oriented testing approach that enables failure detection and attribution by combining (1) adaptive test case generation via seed selection and mutation, (2) capability oracles for identifying capability-specific errors, and (3) a feedback mechanism that attributes failures to capabilities and guides further test generation. Experiments show that our method discovers more failure cases and more accurately pinpoints capability-level deficiencies than state-of-the-art baselines, providing more interpretable and actionable guidance for improving embodied agents.
title Where Did It Go Wrong? Capability-Oriented Failure Attribution for Vision-and-Language Navigation Agents
topic Multiagent Systems
Artificial Intelligence
url https://arxiv.org/abs/2604.25161