Dujun Nie

Xianda Guo, Ruijun Zhang, Yiqun Duan et al.

Accurate spatial reasoning in outdoor environments - covering geometry, object pose, and inter-object relationships - is fundamental to downstream tasks such as mapping, motion forecasting, and high-level planning in autonomous driving. We introduce SURDS, a large-scale benchmark designed to systematically evaluate the spatial reasoning capabilities of vision language models (VLMs). Built on the nuScenes dataset, SURDS comprises 41,080 vision-question-answer training instances and 9,250 evaluation samples, spanning six spatial categories: orientation, depth estimation, pixel-level localization, pairwise distance, lateral ordering, and front-behind relations. We benchmark leading general-purpose VLMs, including GPT, Gemini, and Qwen, revealing persistent limitations in fine-grained spatial understanding. To address these deficiencies, we go beyond static evaluation and explore whether alignment techniques can improve spatial reasoning performance. Specifically, we propose a reinforcement learning-based alignment scheme leveraging spatially grounded reward signals - capturing both perception-level accuracy (location) and reasoning consistency (logic). We further incorporate final-answer correctness and output-format rewards to guide fine-grained policy adaptation. Our GRPO-aligned variant achieves an overall score of 40.80 in the SURDS benchmark. Notably, it outperforms proprietary systems such as GPT-4o (13.30) and Gemini-2.0-flash (35.71). To our best knowledge, this is the first study to demonstrate that reinforcement learning-based alignment can significantly and consistently enhance the spatial reasoning capabilities of VLMs in real-world driving contexts. We release the SURDS benchmark, evaluation toolkit, and GRPO alignment code through: https://github.com/XiandaGuo/Drive-MLLM.

Dujun Nie, Fengjiao Chen, Qi Lv et al.

While the shortage of explicit action data limits Vision-Language-Action (VLA) models, human action videos offer a scalable yet unlabeled data source. A critical challenge in utilizing large-scale human video datasets lies in transforming visual signals into ontology-independent representations, known as latent actions. However, the capacity of latent action representation to derive robust control from visual observations has yet to be rigorously evaluated. We introduce the Latent Action Representation Yielding (LARY) Benchmark, a unified framework for evaluating latent action representations on both high-level semantic actions (what to do) and low-level robotic control (how to do). The comprehensively curated dataset encompasses over one million videos (1,000 hours) spanning 151 action categories, alongside 620K image pairs and 595K motion trajectories across diverse embodiments and environments. Our experiments reveal two crucial insights: (i) General visual foundation models, trained without any action supervision, consistently outperform specialized embodied latent action models. (ii) Latent-based visual space is fundamentally better aligned to physical action space than pixel-based space. These results suggest that general visual representations inherently encode action-relevant knowledge for physical control, and that semantic-level abstraction serves as a fundamentally more effective pathway from vision to action than pixel-level reconstruction.

Xianda Guo, Chenming Zhang, Youmin Zhang et al.

We present LightStereo, a cutting-edge stereo-matching network crafted to accelerate the matching process. Departing from conventional methodologies that rely on aggregating computationally intensive 4D costs, LightStereo adopts the 3D cost volume as a lightweight alternative. While similar approaches have been explored previously, our breakthrough lies in enhancing performance through a dedicated focus on the channel dimension of the 3D cost volume, where the distribution of matching costs is encapsulated. Our exhaustive exploration has yielded plenty of strategies to amplify the capacity of the pivotal dimension, ensuring both precision and efficiency. We compare the proposed LightStereo with existing state-of-the-art methods across various benchmarks, which demonstrate its superior performance in speed, accuracy, and resource utilization. LightStereo achieves a competitive EPE metric in the SceneFlow datasets while demanding a minimum of only 22 GFLOPs and 17 ms of runtime, and ranks 1st on KITTI 2015 among real-time models. Our comprehensive analysis reveals the effect of 2D cost aggregation for stereo matching, paving the way for real-world applications of efficient stereo systems. Code will be available at https://github.com/XiandaGuo/OpenStereo.

Xianda Guo, Chenming Zhang, Youmin Zhang et al.

Stereo matching serves as a cornerstone in 3D vision, aiming to establish pixel-wise correspondences between stereo image pairs for depth recovery. Despite remarkable progress driven by deep neural architectures, current models often exhibit severe performance degradation when deployed in unseen domains, primarily due to the limited diversity of training data. In this work, we introduce StereoAnything, a data-centric framework that substantially enhances the zero-shot generalization capability of existing stereo models. Rather than devising yet another specialized architecture, we scale stereo training to an unprecedented level by systematically unifying heterogeneous stereo sources: (1) curated labeled datasets covering diverse environments, and (2) large-scale synthetic stereo pairs generated from unlabeled monocular images. Our mixed-data strategy delivers consistent and robust learning signals across domains, effectively mitigating dataset bias. Extensive zero-shot evaluations on four public benchmarks demonstrate that Stereo Anything achieves state-of-the-art generalization. This work paves the way towards truly universal stereo matching, offering a scalable data paradigm applicable to any stereo image pair. We extensively evaluate the zero-shot capabilities of our model on four public datasets, showcasing its impressive ability to generalize to any stereo image pair. Code is available at https://github.com/XiandaGuo/OpenStereo.

Dujun Nie, Xianda Guo, Yiqun Duan et al.

Object Goal Navigation-requiring an agent to locate a specific object in an unseen environment-remains a core challenge in embodied AI. Although recent progress in Vision-Language Model (VLM)-based agents has demonstrated promising perception and decision-making abilities through prompting, none has yet established a fully modular world model design that reduces risky and costly interactions with the environment by predicting the future state of the world. We introduce WMNav, a novel World Model-based Navigation framework powered by Vision-Language Models (VLMs). It predicts possible outcomes of decisions and builds memories to provide feedback to the policy module. To retain the predicted state of the environment, WMNav proposes the online maintained Curiosity Value Map as part of the world model memory to provide dynamic configuration for navigation policy. By decomposing according to a human-like thinking process, WMNav effectively alleviates the impact of model hallucination by making decisions based on the feedback difference between the world model plan and observation. To further boost efficiency, we implement a two-stage action proposer strategy: broad exploration followed by precise localization. Extensive evaluation on HM3D and MP3D validates WMNav surpasses existing zero-shot benchmarks in both success rate and exploration efficiency (absolute improvement: +3.2% SR and +3.2% SPL on HM3D, +13.5% SR and +1.1% SPL on MP3D). Project page: https://b0b8k1ng.github.io/WMNav/.