analiza aiqbench
This research introduces CondVLN, a new benchmark designed to diagnose how well vision-language navigation agents handle conditional logic like "if-then" instructions based on real-time environmental observations. By grounding branch conditions in verifiable 3D scene-graph predicates, the study reveals that current state-of-the-art models often fail at logical decision-making even when their basic perception and movement capabilities appear functional. The work also proposes a lightweight neurosymbolic model to separate condition evaluation from navigation execution, offering a clearer path toward building robust embodied AI agents capable of complex reasoning in dynamic spaces.
If, Then, Otherwise: Diagnosing Conditional Branching in Vision-Language Navigation
arXiv:2608.17318v1 Announce Type: cross Abstract: Vision-language navigation agents are often evaluated on their ability to follow route-like instructions toward a fixed goal. Yet, real navigation instructions often depend on observed states of the environment: if a condition holds, then follow one path, otherwise take another. Such instructions require an agent to evaluate scene evidence, select the correct logical branch, and execute the corresponding navigation behavior. Existing evaluations provide limited control over conditional branch execution, making it difficult to determine whether agents fail because of perception, grounding, navigation, or logical decision-making. We introduce CondVLN, a scene-graph-grounded benchmark for diagnosing conditional branching in vision-language navigation. CondVLN programmatically generates instructions whose branch conditions are grounded in verifiable 3D scene-graph predicates, with controlled variation in branch depth, dependency chain length, spatial composition, evidence observability, and instruction horizon. CondVLN contains over 11,500 generated conditional instructions across AI2-THOR, Matterport3D, Gibson, and ReplicaCAD, and evaluates agents using standard VLN metrics and branch-specific diagnostics: Branch Selection Accuracy and Conditional Success Rate. Evaluating four state-of-the-art VLN agents (VLN-Zero, NaVid, NaVILA, and Open-Nav) shows that conditional branching exposes failures that are not captured by standard success rate or path length alone: agents can navigate plausibly while committing to a branch inconsistent with the observed scene condition. We also present a lightweight neurosymbolic branch-selection model that separates condition grounding from navigation execution, improving performance by 2x. CondVLN provides a reusable testbed for measuring whether embodied agents can not only follow instructions, but follow the right instruction under the right condition.
Authors:Seoyoung Lee, Neel P. Bhatt, Pranay Samineni, Cong Liu, S P Sharan, Timothy Barclay, Gregory M. Wagner, Daniel Milan, Sandeep Chinchali, Ufuk Topcu, Atlas Wang
Abstract:Vision-language navigation agents are often evaluated on their ability to follow route-like instructions toward a fixed goal. Yet, real navigation instructions often depend on observed states of the environment: if a condition holds, then follow one path, otherwise take another. Such instructions require an agent to evaluate scene evidence, select the correct logical branch, and execute the corresponding navigation behavior. Existing evaluations provide limited control over conditional branch execution, making it difficult to determine whether agents fail because of perception, grounding, navigation, or logical decision-making. We introduce CondVLN, a scene-graph-grounded benchmark for diagnosing conditional branching in vision-language navigation. CondVLN programmatically generates instructions whose branch conditions are grounded in verifiable 3D scene-graph predicates, with controlled variation in branch depth, dependency chain length, spatial composition, evidence observability, and instruction horizon. CondVLN contains over 11,500 generated conditional instructions across AI2-THOR, Matterport3D, Gibson, and ReplicaCAD, and evaluates agents using standard VLN metrics and branch-specific diagnostics: Branch Selection Accuracy and Conditional Success Rate. Evaluating four state-of-the-art VLN agents (VLN-Zero, NaVid, NaVILA, and Open-Nav) shows that conditional branching exposes failures that are not captured by standard success rate or path length alone: agents can navigate plausibly while committing to a branch inconsistent with the observed scene condition. We also present a lightweight neurosymbolic branch-selection model that separates condition grounding from navigation execution, improving performance by 2x. CondVLN provides a reusable testbed for measuring whether embodied agents can not only follow instructions, but follow the right instruction under the right condition.
11 pages, 1 figure, 3 tables. Project page: this https URL
Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
arXiv:2608.17318 [cs.CV]
(or arXiv:2608.17318v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2608.17318 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 303 cuvinte
Subiecte conexe
Runway’s WorldPrompt and the Engineering of Real-Time Worlds
Latent Space · 25 sept. 2026
Academia is for Ambition — Alex Zhang, MIT
Latent Space · 2 oct. 2026
Anticipate Before Acting: Future-State-Conditioned Vision-Language Navigation
arXiv · 21 iul. 2026
Concrete Evaluations to Investigate the OpenAI Model That Hacked Hugging Face
AI Alignment Forum · 3 aug. 2026