VLN-Zero: Rapid Exploration and Cache-Enabled Neurosymbolic Vision-Language Planning for Zero-Shot Transfer in Robot Navigation

2025-09-22

AI Summary

Existing robot navigation methods exhibit poor generalization in unknown environments, heavily relying on exhaustive exploration or costly parameter fine-tuning. Method: This paper proposes a two-stage zero-shot neuro-symbolic navigation framework. In Stage I, structured prompting guides a vision-language model (VLM) to efficiently explore the environment and construct a symbolic scene graph. In Stage II, neuro-symbolic planning operates over the scene graph, augmented by a cache-reuse mechanism to accelerate decision-making. Contribution/Results: To our knowledge, this is the first zero-shot end-to-end vision-language navigation approach achieving low computational overhead—requiring neither environmental priors nor parameter adaptation. Experiments across diverse unknown environments demonstrate a 2× improvement in success rate over prior zero-shot methods, a 50% reduction in traversal time, and a 55% decrease in VLM invocations—outperforming most fine-tuned baselines.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Planning, Routing, and Scheduling: Planning with Language Models
• Computer Vision: Visual Reasoning & Symbolic Representations

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Rapid adaptation in unseen environments is essential for scalable real-world autonomy, yet existing approaches rely on exhaustive exploration or rigid navigation policies that fail to generalize. We present VLN-Zero, a two-phase vision-language navigation framework that leverages vision-language models to efficiently construct symbolic scene graphs and enable zero-shot neurosymbolic navigation. In the exploration phase, structured prompts guide VLM-based search toward informative and diverse trajectories, yielding compact scene graph representations. In the deployment phase, a neurosymbolic planner reasons over the scene graph and environmental observations to generate executable plans, while a cache-enabled execution module accelerates adaptation by reusing previously computed task-location trajectories. By combining rapid exploration, symbolic reasoning, and cache-enabled execution, the proposed framework overcomes the computational inefficiency and poor generalization of prior vision-language navigation methods, enabling robust and scalable decision-making in unseen environments. VLN-Zero achieves 2x higher success rate compared to state-of-the-art zero-shot models, outperforms most fine-tuned baselines, and reaches goal locations in half the time with 55% fewer VLM calls on average compared to state-of-the-art models across diverse environments. Codebase, datasets, and videos for VLN-Zero are available at: https://vln-zero.github.io/.

Problem

• Addressing poor generalization of navigation policies in unseen environments
• Overcoming computational inefficiency in vision-language navigation methods
• Enabling rapid adaptation without exhaustive exploration or rigid policies

Innovation

• Two-phase framework with exploration and neurosymbolic planning
• Cache-enabled execution reuses computed task-location trajectories
• Vision-language models construct symbolic scene graphs for reasoning