# Aux-Think: Exploring Reasoning Strategies for Data-Efficient Vision-Language Navigation

arXiv: https://arxiv.org/abs/2505.11886 · HTML: https://mlsanity.com/paper/2505.11886/

## Key facts

- Cited by 21 papers in the map (data through 14 Oct 2026).
- Placed in Spatial and embodied reasoning, one of 17 directions in LLM reasoning.
- First posted on arXiv 17 May 2025.

## Placed in

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

## Built on by (most cited first, 21 of 21)

- [MonoDream: Monocular Vision-Language Navigation with Panoramic Dreaming](https://mlsanity.com/paper/2508.02549/) (2025)
- [NavForesee: A Unified Vision-Language World Model for Hierarchical Planning and Dual-Horizon Navigation Prediction](https://mlsanity.com/paper/2512.01550/) (2025)
- [SocialNav: Training Human-Inspired Foundation Model for Socially-Aware Embodied Navigation](https://mlsanity.com/paper/2511.21135/) (2025)
- [VLingNav: Embodied Navigation with Adaptive Reasoning and Visual-Assisted Linguistic Memory](https://mlsanity.com/paper/2601.08665/) (2026)
- [Vision to Geometry: 3D Spatial Memory for Sequential Embodied MLLM Reasoning and Exploration](https://mlsanity.com/paper/2512.02458/) (2025)
- [D3D-VLP: Dynamic 3D Vision-Language-Planning Model for Embodied Grounding and Navigation](https://mlsanity.com/paper/2512.12622/) (2025)
- [textscNaVIDA: Vision-Language Navigation with Inverse Dynamics Augmentation](https://mlsanity.com/paper/2601.18188/) (2026)
- [MapDream: Task-Driven Map Learning for Vision-Language Navigation](https://mlsanity.com/paper/2602.00222/) (2026)
- [GA-VLN: Geometry-Aware BEV Representation for Efficient Vision-Language Navigation](https://mlsanity.com/paper/2605.22036/) (2026)
- [SVLL: Staged Vision-Language Learning for Physically Grounded Embodied Task Planning](https://mlsanity.com/paper/2603.11563/) (2026)
- [HoloAgent-0: A Unified Embodied Agent Framework with 3D Spatial Memory](https://mlsanity.com/paper/2606.23565/) (2026)
- [A Comprehensive Survey and Systematic Real-World Evaluation of Embodied Vision-and-Language Navigation](https://mlsanity.com/paper/2607.09792/) (2026)
- [SC2-WM: A Self-Correcting World Model with Closed-Loop Feedback for Vision-and-Language Navigation in Continuous Environments](https://mlsanity.com/paper/2608.07548/) (2026)
- [LightNav-0: Eliciting VLM Spatial Intelligence for Generalist Embodied Navigation](https://mlsanity.com/paper/2608.30935/) (2026)
- [HiRO-Nav: Hybrid ReasOning Enables Efficient Embodied Navigation](https://mlsanity.com/paper/2604.08232/) (2026)
- [NavOne: One-Step Global Planning for Vision-Language Navigation on Top-Down Maps](https://mlsanity.com/paper/2605.06317/) (2026)
- [Traj-VLN: Learning Pixel-Space Interaction via Autoregressive Trajectory Generation](https://mlsanity.com/paper/2607.10744/) (2026)
- [PanoVLN: Towards Effective Panoramic Vision-and-Language Navigation](https://mlsanity.com/paper/2609.34759/) (2026)
- [Seek Before You Move: Evidence Seeking for Progress Grounding in Vision-Language Navigation](https://mlsanity.com/paper/2609.37353/) (2026)
- [StageVLN: Spatial and Trajectory Auxiliary Guidance for Efficient Vision-Language Navigation](https://mlsanity.com/paper/2610.05664/) (2026)
- [GuidedNav: Shaping Vision-Language Navigation Representations through Pre-action Attention and Spatial Guidance](https://mlsanity.com/paper/2610.13122/) (2026)

## Abstract (arXiv)

Vision-Language Navigation (VLN) is a critical task for developing embodied agents that can follow natural language instructions to navigate in complex real-world environments. Recent advances in VLN by large pretrained models have significantly improved generalization and instruction grounding compared to traditional approaches. However, the role of reasoning strategies in navigation-an action-centric, long-horizon task-remains underexplored, despite Chain-of-Thought (CoT) reasoning's demonstrated success in static tasks like visual question answering. To address this gap, we conduct the first systematic evaluation of reasoning strategies for VLN, including No-Think (direct action prediction), Pre-Think (reason before action), and Post-Think (reason after action). Surprisingly, our findings reveal the Inference-time Reasoning Collapse issue, where inference-time reasoning degrades navigation accuracy, highlighting the challenges of integrating reasoning into VLN. Based on this insight, we propose Aux-Think, a framework that trains models to internalize structured reasoning patterns through CoT supervision, while inferring action directly without reasoning in online prediction. To support this framework, we release R2R-CoT-320k, the first Chain-of-Thought annotated dataset for VLN. Extensive experiments show that Aux-Think reduces training effort greatly and achieves the best performance under the same data scale.