This work proposes an alternative approach: fine-tuning VLMs to learn navigation interactions directly in 2D pixel space through autoregressive trajectory generation, and demonstrates that this flagship model achieves state-of-the-art level performance with relatively limited computational resources and training data.
Abstract
Benefiting from the powerful priors embedded in large-scale pre-training data and the emerging commonsense reasoning ability, large language models (LLMs) have shown unprecedented generalization capabilities in many research fields. Recently, projecting visual embeddings into the language space via vision-language models (VLMs) to achieve sim-toreal and cross-scene generalization has become a prevailing paradigm in the field of Vision-and-Language Navigation in Continuous Environments (VLN-CE). VLN requires an embodied agent to navigate through unseen environments following natural linguistic instructions. We emphasize that a VLN task can be decomposed into a sequence of sub-tasks, each corresponding to a process of 3D spatial interaction with the environments described by instructions such as"walk to the end of the sofa and turn left."However, such spatial interactions involving moving into the image along the direction of depth sensing are puzzling for VLMs as they were predominantly trained on conversations with RGB images. Rather than incorporating depth or 3D geometric information-which VLMs rarely encounter during pretrainingwe propose an alternative approach: fine-tuning VLMs to learn navigation interactions directly in 2D pixel space through autoregressive trajectory generation. Given a linguistic instruction and historical observations, our model sequentially predicts a series of pixel coordinates, drawing a trajectory from the bottom center of the current observation. While prior work has proved that pixel-goal supervision outperforms learning of discrete actions, our experiments further verify that the supervision of pixel-space trajectory significantly enhances VLN performance. Moreover, we demonstrate that our flagship model achieves state-of-the-art level performance with relatively limited computational resources and training data.
The S-squared-VLA is proposed, which explicitly decouples the semantic and spatial streams in Vision-Language-Action models, and significantly outperforms baselines, achieving the highest No Collision rate among all evaluated methods.
Recent vision-language navigation (VLN) systems increasingly adapt pretrained vision-language models (VLMs) into vision-language-action (VLA) policies that map egocentric observations and language instructions directly to navigation actions. Although semantically capable, such action-centric training does not explicitly model how the agent's visual observations should evolve under its predicted motion. Generative world-action models (WAMs) jointly predict future observations and actions, yet existing WAMs for continuous VLN do not condition joint future-view and action generation on geometry-aware representations inferred from the observed history. We present WNM-3D, a generative World Navigation Model with 3D scene conditioning for continuous VLN. To consolidate past observations into persistent scene context, a frozen feed-forward geometry encoder extracts geometry-aware representations from the monocular egocentric RGB history, and a trainable 3D Scene-to-Token Adapter converts them into a fixed-length prefix in the token space of the world-action Diffusion Transformer. Through block-causal attention, this prefix conditions every future video-action block, providing a shared geometric context for both future-view and action generation. We train WNM-3D through supervised world-action fine-tuning on A*-generated demonstrations, DAgger-style adaptation on policy-visited states, and Counterfactual DanceGRPO refinement for closed-loop execution. Experiments on GN-Bench show that WNM-3D outperforms strong VLM-based navigation policies and its 2D-conditioned counterpart in closed-loop navigation. Stage-wise ablations further show that DAgger-SFT provides the larger success-rate gain, while Counterfactual DanceGRPO subsequently improves both navigation success and path efficiency.
Yuehao Huang, Yunzi Wu, Xiaotao Zhang et al.· 0 citations
HAM-VLN is presented, a decision-coupled, agent-authored memory that equips the robot with a persistent, depth-grounded world graph and reduces the context length by more than 65% compared to previous methods.
An Liu, Bingxi Liu, Hongyu Ding et al.· arXiv.org· 0 citations
Recent vision-language-action (VLA) methods improve manipulation performance by aligning their representations with 3D scene geometry. However, these methods often struggle with long-horizon manipulation and observation aliasing between visually similar states due to a lack of temporal information: the 3D scene geometry captures only the current state, rather than how it has evolved over time. To resolve this, we present Temporal Forcing, a 4D representation alignment method for VLA models. Specifically, we first introduce a history pathway that enables a vanilla VLA model to summarize observation history into temporally aware latent representations. Then, the latent representations are aligned with the geometric features extracted by a pretrained 4D foundation model, which captures the evolving 3D world through temporally consistent geometric representations, enabling a deeper understanding of dynamic environments. Temporal Forcing reaches 98.8% on LIBERO, outperforming its base model by 2.2 points. On a physical hidden-placement task, it raises full-task success from 20.0% to 43.3%. Code will be publicly available.
Xing-Yu Ding, Yuzhong Zhao, Chunming Zhao et al.· 0 citations
Space Tokens is introduced, a lightweight, architecture-agnostic framework that equips VLMs with explicit continuous spatial representations without requiring additional inference-time modules, and demonstrates that continuous spatial tokens provide an effective, interpretable, and computationally efficient mechanism for integrating geometric reasoning into large vision-language models.
Hunter Schofield, Mohammed Elmahgiubi, Mohammad Mahdavian et al.· 0 citations
GaussianWAM is proposed, a training-time representation-enhancement framework that organizes geometric and semantic supervision through a 3D Gaussian field and improves performance on standard LIBERO and shows positive transfer trends on RoboTwin and real-world manipulation.
Zijian Zhang, Yuqing Jiang, Weitao Zhou et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
