GN0: Toward a Unified Paradigm for Generation, Evaluation, and Policy Learning in Visual-Language Navigation
For embodied navigation researchers, this work provides a comprehensive framework (data, simulation, benchmark, and learning) that improves generalization and long-horizon capabilities, though it is incremental as it combines existing techniques (3DGS, RL, DAgger) rather than introducing a fundamentally new paradigm.
The paper tackles data scarcity and limited generalization in Vision-and-Language Navigation (VLN) by introducing a unified paradigm (GN0) that includes a large-scale dataset (GN-Matrix), a high-fidelity simulation platform based on 3D Gaussian Splatting, a BEV-based benchmark (GN-Bench) with dynamic avatars, and an RL-driven navigation model (BAE). GN0 outperforms state-of-the-art VLN methods on GN-Bench and VLN-CE.
Embodied navigation connects intelligent agents with the physical world and is fundamental for general robotic intelligence. Limited availability and quality of navigation data have constrained Vision-and-Language Navigation (VLN) systems' generalization and long-horizon capabilities. To address this, we curate diverse 3D scenes and develop an automated pipeline for large-scale navigation data, resulting in the GN-Matrix dataset. Building on a 3D Gaussian Splatting (3DGS) engine, we introduce a high-fidelity simulation platform supporting interactive roaming and collision-aware navigation. We further propose GN-Bench, the first BEV-based benchmark incorporating dynamic 3DGS avatars for human-robot interaction evaluation. To leverage the simulator, we develop an RL-driven navigation foundation model, Break and Establish (BAE). After supervised learning, DAgger exposes the model to rollout-induced states, breaking narrow expert-centric distributions and enabling downstream RL exploration. This unified VLN paradigm integrates map-based and map-free tasks, including instruction following, human following, and goal navigation. GN-BAE formalizes high-fidelity 3DGS-rendered Bird's Eye View representations as compact memory, unlocking latent spatial reasoning in VLMs. Extensive evaluations on GN-Bench and VLN-CE show that GN0 outperforms state-of-the-art VLN methods. Overall, GN-Matrix offers a unified framework spanning data, simulation, and learning, advancing embodied navigation in research and industrial applications.