This analysis reveals a gap between benchmark-oriented spatial specialization and navigation performance, and shows how aligning spatial supervision with navigation goals, phases, and decision learning improves navigation.
Abstract
Does progress on spatial reasoning benchmarks translate into better navigation? Existing benchmarks test isolated inferences from images or videos, with little connection to downstream navigation. Our analysis reveals a gap between benchmark-oriented spatial specialization and navigation performance, and shows how aligning spatial supervision with navigation goals, phases, and decision learning improves navigation. Guided by these findings, we build \textsc{Spatial-Nav-100K} and fine-tune in two stages, \textit{i.e.} first learning a shared spatial-navigation foundation, and then specializing each phase with the abilities it relies on. We further introduce Spatial-NPD, where a teacher conditioned on spatial priors produces grounded action preferences and distills them into a student policy, so no explicit spatial reasoning is needed at inference. With 45 A100 GPU-hours of policy training, our 8B model reaches SR/SPL of 77.4/35.4 on HM3D-v0.2, 60.2/30.5 on HM3D-v0.1, and 47.9/20.6 on train-unseen MP3D. It outperforms several systems that rely on closed-source models or thousands of GPU-hours of training, at 148 ms per action step. All code and datasets will be publicly available at https://github.com/ylwhxht/Spatial-Nav.
VCN-Bench is introduced, a VCN-Bench benchmark for probing closed-loop spatial reasoning over prior visual experience in MLLMs, and MV-DualVLN is proposed, a planning-oriented baseline that jointly leverages prior video and in-episode observations.
Si-Qi Zhang, Meng Wei, Chen-Yang Wan et al.· arXiv.org· 0 citations
Language-guided navigation requires connecting partial observations to a persistent spatial reference and learning how actions change that representation. We introduce InsightMap, a framework that uses top-down maps as both explicit spatial memory and action-conditioned prediction targets. Historical views are linked t...
Existing spatial reasoning benchmarks mainly test spatial perception: reading off relations already visible in the input. Yet real-world spatial intelligence demands predictive spatial reasoning: constructing a scene from observations, anticipating how an intervention changes it, and reasoning about the unseen outcome....
Hong-Xing Li, Jin-Yue Su, Ding-Ming Li et al.· 1 citation
Spatial reasoning benchmarks evaluate vision-language models across diverse tasks, but task-level scores do not reveal which underlying capabilities account for success or failure. Each task requires recovering spatial evidence, representing geometry, and reasoning over it. We disentangle these capabilities by comparin...
Peng-Zhan Sun, Jun-Bin Xiao, Ramanathan Rajaraman et al.· 0 citations
SpatialMind is introduced, a metric-scale VLM for spatial reasoning and future prediction that substantially outperforms both general and spatially specialized models on the authors' benchmark while achieving competitive zero-shot performance on VSI-Bench, OSI-Bench, and VLM4D.
Fei-Ran Wang, Xiao-Qi Wang, Zi-Wei Li et al.· 0 citations
This article proposes ReasonWalker, a novel navigation model designed to enable reasoning-based navigation using implicit instructions over time, and proposes a new hierarchical learning paradigm, where the model first learns navigation actions and then acquires scene associations for implicit instruction reasoning.
Xudong Wang, Bai-Cheng Liu, Jia-Hua Dong et al.· IEEE Transactions on Cyberne...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.