# MSNav: Zero-Shot Vision-and-Language Navigation with Dynamic Memory and LLM Spatial Reasoning

arXiv: https://arxiv.org/abs/2508.16654 · HTML: https://mlsanity.com/paper/2508.16654/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- Builds on 17 of its 51 references that are in the map.
- First posted on arXiv 20 Aug 2025.

## Placed in

- LLM reasoning / [Spatial and embodied reasoning](https://mlsanity.com/lr/spatial-and-embodied-reasoning/)

## Builds on (most cited first, 17 of 17)

- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [GPT-4o System Card](https://mlsanity.com/paper/2410.21276/) (2024)
- [Large Language Models are Zero-Shot Reasoners](https://mlsanity.com/paper/2205.11916/) (2022)
- [YOLO-World: Real-Time Open-Vocabulary Object Detection](https://mlsanity.com/paper/2401.17270/) (2024)
- [NavGPT: Explicit Reasoning in Vision-and-Language Navigation with Large Language Models](https://mlsanity.com/paper/2305.16986/) (2023)
- [Think Global, Act Local: Dual-scale Graph Transformer for Vision-and-Language Navigation](https://mlsanity.com/paper/2202.11742/) (2022)
- [Towards Learning a Generalist Model for Embodied Navigation](https://mlsanity.com/paper/2312.02010/) (2023)
- [Airbert: In-domain Pretraining for Vision-and-Language Navigation](https://mlsanity.com/paper/2108.09105/) (2021)
- [Discuss Before Moving: Visual Language Navigation via Multi-expert Discussions](https://mlsanity.com/paper/2309.11382/) (2023)
- [NavGPT-2: Unleashing Navigational Reasoning Capability for Large Vision-Language Models](https://mlsanity.com/paper/2407.12366/) (2024)
- [NavCoT: Boosting LLM-Based Vision-and-Language Navigation via Learning Disentangled Reasoning](https://mlsanity.com/paper/2403.07376/) (2024)
- [Target-Driven Structured Transformer Planner for Vision-Language Navigation](https://mlsanity.com/paper/2207.11201/) (2022)
- [Learning from Unlabeled 3D Environments for Vision-and-Language Navigation](https://mlsanity.com/paper/2208.11781/) (2022)
- [Scene-Intuitive Agent for Remote Embodied Visual Grounding](https://mlsanity.com/paper/2103.12944/) (2021)
- [LangNav: Language as a Perceptual Representation for Navigation](https://mlsanity.com/paper/2310.07889/) (2023)

## Built on by (most cited first, 4 of 4)

- [ProFocus: Proactive Perception and Focused Reasoning in Vision-and-Language Navigation](https://mlsanity.com/paper/2603.05530/) (2026)
- [Does Peer Observation Help? Vision-Sharing Collaboration for Vision-Language Navigation](https://mlsanity.com/paper/2603.20804/) (2026)
- [Global Commander and Local Operative: A Dual-Agent Framework for Scene Navigation](https://mlsanity.com/paper/2602.18941/) (2026)
- [Reasmory: 3D Reconstruction as Explicit Memory for VLMs Spatial Reasoning](https://mlsanity.com/paper/2606.00963/) (2026)

## Abstract (arXiv)

Vision-and-Language Navigation (VLN) requires an agent to interpret natural language instructions and navigate complex environments. Current approaches often adopt a "black-box" paradigm, where a single Large Language Model (LLM) makes end-to-end decisions. However, it is plagued by critical vulnerabilities, including poor spatial reasoning, weak cross-modal grounding, and memory overload in long-horizon tasks. To systematically address these issues, we propose Memory Spatial Navigation(MSNav), a framework that fuses three modules into a synergistic architecture, which transforms fragile inference into a robust, integrated intelligence. MSNav integrates three modules: Memory Module, a dynamic map memory module that tackles memory overload through selective node pruning, enhancing long-range exploration; Spatial Module, a module for spatial reasoning and object relationship inference that improves endpoint recognition; and Decision Module, a module using LLM-based path planning to execute robust actions. Powering Spatial Module, we also introduce an Instruction-Object-Space (I-O-S) dataset and fine-tune the Qwen3-4B model into Qwen-Spatial (Qwen-Sp), which outperforms leading commercial LLMs in object list extraction, achieving higher F1 and NDCG scores on the I-O-S test set. Extensive experiments on the Room-to-Room (R2R) and REVERIE datasets demonstrate MSNav's state-of-the-art performance with significant improvements in Success Rate (SR) and Success weighted by Path Length (SPL).