Data-Centric Recursive Self-Improvement for Navigation via Policy-Synchronized Language-Action Dual Bootstrapping

Songze Li; Haoming Gu; Ziyang Gong; Zhitong Wang; Gengze Zhou; Zun Wang; Yuhan Zhu; Tianxiang Jiang

cs.CVOct 11, 2026

Data-Centric Recursive Self-Improvement for Navigation via Policy-Synchronized Language-Action Dual Bootstrapping

Authors: Songze Li, Haoming Gu, Ziyang Gong, Zhitong Wang, Gengze Zhou, Zun Wang, Yuhan Zhu, Tianxiang Jiang, +6 more

Organizations: Fudan University · Shanghai AI Laboratory · Harbin Institute of Technology · Shanghai Jiao Tong University · Tsinghua University · University of Adelaide · UNC Chapel Hill · Nanjing University · University of Science and Technology of China

Abstract

Vision-language navigation requires agents to follow natural-language instructions to reach target locations, making high-quality language supervision essential. Since human annotation is costly, prior work has explored scalable data synthesis methods. However, existing methods typically either use separate models for navigation and instruction generation or couple the two tasks through task-specific components; most also annotate the full trajectory corpus with a fixed model before updating, which can cause generated supervision to lag behind the evolving policy. Therefore, we introduce RSI-Nav, a data-centric recursive self-improvement framework for navigation via policy-synchronized language-action dual bootstrapping, which unifies vision-language-to-action instruction following and vision-action-to-language instruction generation within a shared model. Specifically, after supervised fine-tuning establishes fundamental capabilities, RSI-Nav progressively processes unlabeled trajectories in small shards. The current model generates instructions for these trajectories, executes them to evaluate goal completion and path efficiency, retains qualified samples, and immediately updates itself before processing the next shard. This shard-wise update keeps newly generated supervision aligned with the evolving policy. We also incorporate spatial understanding data, since stronger spatial reasoning benefits navigation. Extensive experiments show that RSI-Nav reaches 68.6% SR on R2R-CE and 72.4% SR on RxR-CE, while providing stronger instruction-generation utility and modestly improved spatial understanding after recursive evolution.

Figures & tables

Figure 1: RSI-Nav unifies navigation, instruction generation, and spatial understanding in one model, and recursively improves through policy-synchronized language-action dual bootstrapping.

Figure 2: Comparison of three VLN training paradigms. (a) Dual-task learning relies on task-specific architectures and costly human annotation. (b) Data flywheel refines data at scale with separate navigation and instruction-generation models. (c) RSI-Nav generates and filters data with the current policy and updates immediately after each shard, forming a policy-synchronized language-action dual bootstrapping loop for data-centric recursive self-improvement.

Figure 3: Overview of RSI-Nav. The current policy generates instructions for unlabeled trajectories and executes them to evaluate goal completion and efficiency. Qualified instructions are paired with the original trajectories to train instruction following and instruction generation jointly. The model is then updated with the accepted samples and spatial replay data before processing the next shard.

Method	Observation				Spatial Understanding				R2R-CE Val-Unseen				RxR-CE Val-Unseen
Method	S.RGB	Pano.	Depth	Odo.	VSI-Bench $\uparrow$	ScanQA $\uparrow$	SQA3D $\uparrow$	MindCube $\uparrow$	NE $\downarrow$	OSR $\uparrow$	SR $\uparrow$	SPL $\uparrow$	NE $\downarrow$	SR $\uparrow$	SPL $\uparrow$	nDTW $\uparrow$
Waypoint-Predictor-Based VLN Models
VLN $\circlearrowright$ BERT ∗ ( Hong et al., 2022 )		✓	✓	✓	-	-	-	-	5.74	53.0	44.0	39.0	8.98	27.0	22.6	46.7
Sim2Sim ∗ ( Krantz and Lee, 2022 )		✓	✓	✓	-	-	-	-	6.07	52.0	43.0	36.0	-	-	-	-
GridMM ∗ ( Wang et al., 2023b )		✓	✓	✓	-	-	-	-	5.11	61.0	49.0	41.0	-	-	-	-
Reborn ∗ ( An et al., 2022 )		✓	✓	✓	-	-	-	-	5.40	57.0	50.0	46.0	5.98	48.6	42.0	63.3

Table 1: Comparison on R2R-CE and RxR-CE Val-Unseen and four spatial-understanding benchmarks. ∗ denotes methods using the waypoint predictor from Hong et al. (2022) .

Table 5

Figure 4: Qualitative results of RSI-Nav in real-world environments.

Method	Instruction Generation						Training Utility
Method	BLEU-1 $\uparrow$	BLEU-4 $\uparrow$	METEOR $\uparrow$	ROUGE $\uparrow$	CIDEr $\uparrow$	SPICE $\uparrow$	NE $\downarrow$	OSR $\uparrow$	SR $\uparrow$	SPL $\uparrow$
BT-speaker ( Fried et al., 2018 )	68.2	26.3	21.7	48.0	37.9	18.9	-	-	-	-
EnvDrop ( Tan et al., 2019 )	72.3	27.1	23.6	49.0	41.7	21.8	-	-	-	-
CCC ( Wang et al., 2022 )	70.8	27.2	23.1	47.7	46.1	21.4	-	-	-	-
KEFA ( Zeng et al., 2023 )	73.8	28.3	24.4	50.3	42.7	23.4	-	-	-	-
LANA ( Wang et al., 2023a )	73.6	28.9	23.7	49.8	45.7	22.6	-	-	-	-

Table 4: Trajectory-to-instruction generation and training utility. We report some NLP metrics for instruction generation and evaluate training utility by training navigation models on the generated supervision and measuring downstream R2R-CE navigation performance.

Training Objective		Instruction Following (VL2A)				Instruction Generation (VA2L)
VL2A	VA2L	NE $\downarrow$	OSR $\uparrow$	SR $\uparrow$	SPL $\uparrow$	BLEU-1 $\uparrow$	BLEU-4 $\uparrow$	METEOR $\uparrow$	ROUGE $\uparrow$	CIDEr $\uparrow$	SPICE $\uparrow$
✓		6.19	57.2	44.7	38.6	-	-	-	-	-	-
	✓	-	-	-	-	71.7	28.8	23.3	48.8	49.8	20.6
✓	✓	5.69	59.9	48.8	41.0	71.0	28.2	23.2	48.5	50.6	20.1

Table 5: Ablation of bidirectional language–action learning. Joint VL2A–VA2L training substantially improves navigation while maintaining comparable instruction-generation quality.

Appendix figures & tables4 assets

Supplementary material from the paper’s appendix.

Appendix

Representation	NE $\downarrow$	OSR $\uparrow$	SR $\uparrow$	SPL $\uparrow$
Arrow	6.32	51.22	43.28	37.80
Single	5.56	57.48	48.72	42.67
Double	5.92	53.78	43.83	39.22

Appendix

Table A1: Action representation ablation on R2R val-unseen.

Representation	Supervision	Sampling	NE $\downarrow$	OSR $\uparrow$	SR $\uparrow$	SPL $\uparrow$
Arrow	V2L	Interval-4	6.39	48.23	41.44	36.27
Arrow	VA2L	Interval-4	6.20	47.14	41.49	37.62
Arrow	V2L	Uniform-32	6.41	48.67	42.09	36.82
Arrow	VA2L	Uniform-32	5.71	50.68	44.15	39.54
Single	V2L	Interval-4	5.40	54.98	47.25	40.59
Single	VA2L	Interval-4	5.42	55.36	48.02	41.36

Appendix

Table A2: Downstream navigation results with different action inputs and frame sampling strategies.

Figure 5: Selected correct QA examples on VSI-Bench, ScanQA, SQA3D, and MindCube.

Figure 6: Visualization of navigation trajectories on R2R-CE and RxR-CE.

Explore similar work

GuidedNav: Shaping Vision-Language Navigation Representations through Pre-action Attention and Spatial Guidance
Oct 9, 2026Jingyu Guo, Jiaxin Huang, Chengxing Lin +6Vision-Language Navigation Visual Navigation
University of Melbourne · AgiBot · Mohamed bin Zayed University of Artificial Intelligence +3
AwareVLN: Reasoning with Self-awareness for Vision-Language Navigation
May 21, 2026Wenxuan Guo, Xiuwei Xu, Yichen Liu +7Vision-Language Navigation VLM Reasoning
Tsinghua University, China
Path-level Hindsight Instructions for Semantic Exploration in Vision-Language Navigation
Jul 2, 2026Sung June Kim, Sangpil Kim, Honglak LeeEmbodied Navigation Vision-Language Navigation
Korea University, Seoul, South Korea · University of Michigan, Ann Arbor, Michigan, USA

Oct 9, 2026J/K move · Enter open · S save

GuidedNav: Shaping Vision-Language Navigation Representations through Pre-action Attention and Spatial Guidance

Jingyu Guo, Jiaxin Huang, Chengxing Lin, Xiaoyu Luo+5

University of Melbourne · AgiBot · Mohamed bin Zayed University of Artificial Intelligence+3

Vision-Language Navigation Visual Navigation

Open paper AskPDF arXiv

Vision-and-language navigation (VLN) requires an agent to track progress through an extended instruction, identify landmarks relevant to its current objective, and understand the surrounding spatial structure to act. End-to-end navigation policies typically rely on action supervision to implicitly learn these capabilities. Without explicit guidance on the evidence underlying each decision, however, policies can overfit to correlations predictive of training actions rather than learn the task-relevant grounding and geometry needed for generalization. Our insight is to guide the representations that connect instructions to actions: identifying what remains to be done determines which visual evidence matters, while understanding the geometry around that evidence informs how to move. We introduce GuidedNav, a unified framework that explicitly shapes these representations through training-time guidance. Subtask and landmark guidance steer native pre-action attention toward the active instruction segment and relevant visual regions, grounding action prediction in the current objective. Spatial guidance aligns the policy's internal representations with geometric features that capture long-horizon context and local spatial structure. Together, these objectives strengthen navigation capabilities while avoiding the inference overhead of intermediate textual reasoning and online geometric encoding. We also introduce Guided-R2R and Guided-RxR, two annotation-enriched navigation datasets that support more interpretable VLN research. GuidedNav establishes state-of-the-art performance among the compared monocular RGB policies, improving success rates by 8.1% and 13.1% relative to the strongest prior results on R2R-CE and RxR-CE val-unseen, respectively, while achieving faster inference. Extensive experiments further demonstrate the effectiveness and real-world transferability of our approach.

Abstract

Explore similar work

GuidedNav: Shaping Vision-Language Navigation Representations through Pre-action Attention and Spatial Guidance

AwareVLN: Reasoning with Self-awareness for Vision-Language Navigation

Path-level Hindsight Instructions for Semantic Exploration in Vision-Language Navigation