EmbodiedMidtrain: Bridging the Gap between Vision-Language Models and Vision-Language-Action Models via Mid-training

Vision-Language-Action Models (VLAs) inherit their visual and linguistic capabilities from Vision-Language Models (VLMs), yet most VLAs are built from off-the-shelf VLMs that are not adapted to the embodied domain, limiting their downstream performance. In this work, we propose EmbodiedMidtrain to bridge the gap between VLMs and VLAs. We first characterize the data distribution gap between them, showing that VLA data occupy compact regions that are largely separated from the broader VLM distribution, while the degree of alignment varies substantially both across and within VLM data sources. Then, we build a mid-training data engine that leverages a lightweight learnable proximity estimator to select the most VLA-aligned candidates from a large VLM pool, and mid-trains the VLM on this curated mixture before downstream VLA fine-tuning. Experiments on three robot manipulation benchmarks show that mid-training consistently improves performance across different VLM backbones, achieving results competitive with expert VLAs and off-the-shelf VLMs trained with larger model scale and training budgets. Further analysis reveals that mid-training provides a stronger initialization for VLA fine-tuning, with gains emerging from the earliest steps and widening throughout training. Moreover, the data engine captures both dataset-level and sample-level alignment signals, favoring spatial reasoning over text-centric tasks while preserving the diversity of the VLM data. We will release all code, data and models for future research.

ExploreVLA: Dense World Modeling and Exploration for End-to-End Autonomous Driving

End-to-end autonomous driving models based on Vision-Language-Action (VLA) architectures have shown promising results by learning driving policies through behavior cloning on expert demonstrations. However, imitation learning inherently limits the model to replicating observed behaviors without exploring diverse driving strategies, leaving it brittle in novel or out-of-distribution scenarios. Reinforcement learning (RL) offers a natural remedy by enabling policy exploration beyond the expert distribution. Yet VLA models, typically trained on offline datasets, lack directly observable state transitions, necessitating a learned world model to anticipate action consequences. In this work, we propose a unified understanding-and-generation framework that leverages world modeling to simultaneously enable meaningful exploration and provide dense supervision. Specifically, we augment trajectory prediction with future RGB and depth image generation as dense world modeling objectives, requiring the model to learn fine-grained visual and geometric representations that substantially enrich the planning backbone. Beyond serving as a supervisory signal, the world model further acts as a source of intrinsic reward for policy exploration: its image prediction uncertainty naturally measures a trajectory's novelty relative to the training distribution, where high uncertainty indicates out-of-distribution scenarios that, if safe, represent valuable learning opportunities. We incorporate this exploration signal into a safety-gated reward and optimize the policy via Group Relative Policy Optimization (GRPO). Experiments on the NAVSIM and nuScenes benchmarks demonstrate the effectiveness of our approach, achieving a state-of-the-art PDMS score of 93.7 and an EPDMS of 88.8 on NAVSIM. The code and demo will be publicly available at https://zihaosheng.github.io/ExploreVLA/.

UniDAC: Universal Metric Depth Estimation for Any Camera

Monocular metric depth estimation (MMDE) is a core challenge in computer vision, playing a pivotal role in real-world applications that demand accurate spatial understanding. Although prior works have shown promising zero-shot performance in MMDE, they often struggle with generalization across diverse camera types, such as fisheye and $360^\circ$ cameras. Recent advances have addressed this through unified camera representations or canonical representation spaces, but they require either including large-FoV camera data during training or separately trained models for different domains. We propose UniDAC, an MMDE framework that presents universal robustness in all domains and generalizes across diverse cameras using a single model. We achieve this by decoupling metric depth estimation into relative depth prediction and spatially varying scale estimation, enabling robust performance across different domains. We propose a lightweight Depth-Guided Scale Estimation module that upsamples a coarse scale map to high resolution using the relative depth map as guidance to account for local scale variations. Furthermore, we introduce RoPE-$φ$, a distortion-aware positional embedding that respects the spatial warping in Equi-Rectangular Projections (ERP) via latitude-aware weighting. UniDAC achieves state of the art (SoTA) in cross-camera generalization by consistently outperforming prior methods across all datasets.

Reconstruction Matters: Learning Geometry-Aligned BEV Representation through 3D Gaussian Splatting

Bird's-Eye-View (BEV) perception serves as a cornerstone for autonomous driving, offering a unified spatial representation that fuses surrounding-view images to enable reasoning for various downstream tasks, such as semantic segmentation, 3D object detection, and motion prediction. However, most existing BEV perception frameworks adopt an end-to-end training paradigm, where image features are directly transformed into the BEV space and optimized solely through downstream task supervision. This formulation treats the entire perception process as a black box, often lacking explicit 3D geometric understanding and interpretability, leading to suboptimal performance. In this paper, we claim that an explicit 3D representation matters for accurate BEV perception, and we propose Splat2BEV, a Gaussian Splatting-assisted framework for BEV tasks. Splat2BEV aims to learn BEV feature representations that are both semantically rich and geometrically precise. We first pre-train a Gaussian generator that explicitly reconstructs 3D scenes from multi-view inputs, enabling the generation of geometry-aligned feature representations. These representations are then projected into the BEV space to serve as inputs for downstream tasks. Extensive experiments on nuScenes and argoverse dataset demonstrate that Splat2BEV achieves state-of-the-art performance and validate the effectiveness of incorporating explicit 3D reconstruction into BEV perception.

No Calibration, No Depth, No Problem: Cross-Sensor View Synthesis with 3D Consistency

We present the first study of cross-sensor view synthesis across different modalities. We examine a practical, fundamental, yet widely overlooked problem: getting aligned RGB-X data, where most RGB-X prior work assumes such pairs exist and focuses on modality fusion, but it empirically requires huge engineering effort in calibration. We propose a match-densify-consolidate method. First, we perform RGB-X image matching followed by guided point densification. Using the proposed confidence-aware densification and self-matching filtering, we attain better view synthesis and later consolidate them in 3D Gaussian Splatting (3DGS). Our method uses no 3D priors for X-sensor and only assumes nearly no-cost COLMAP for RGB. We aim to remove the cumbersome calibration for various RGB-X sensors and advance the popularity of cross-sensor learning by a scalable solution that breaks through the bottleneck in large-scale real-world RGB-X data collection.

UniDrive-WM: Unified Understanding, Planning and Generation World Model For Autonomous Driving

World models have become central to autonomous driving, where accurate scene understanding and future prediction are crucial for safe control. Recent work has explored using vision-language models (VLMs) for planning, yet existing approaches typically treat perception, prediction, and planning as separate modules. We propose UniDrive-WM, a unified VLM-based world model that jointly performs driving-scene understanding, trajectory planning, and trajectory-conditioned future image generation within a single architecture. UniDrive-WM's trajectory planner predicts a future trajectory, which conditions a VLM-based image generator to produce plausible future frames. These predictions provide additional supervisory signals that enhance scene understanding and iteratively refine trajectory generation. We further compare discrete and continuous output representations for future image prediction, analyzing their influence on downstream driving performance. Experiments on the challenging Bench2Drive benchmark show that UniDrive-WM produces high-fidelity future images and improves planning performance by 5.9% in L2 trajectory error and 9.2% in collision rate over the previous best method. These results demonstrate the advantages of tightly integrating VLM-driven reasoning, planning, and generative world modeling for autonomous driving. The project page is available at https://unidrive-wm.github.io/UniDrive-WM .

Open Ad-hoc Categorization with Contextualized Feature Learning

Adaptive categorization of visual scenes is essential for AI agents to handle changing tasks. Unlike fixed common categories for plants or animals, ad-hoc categories are created dynamically to serve specific goals. We study open ad-hoc categorization: Given a few labeled exemplars and abundant unlabeled data, the goal is to discover the underlying context and to expand ad-hoc categories through semantic extension and visual clustering around it. Building on the insight that ad-hoc and common categories rely on similar perceptual mechanisms, we propose OAK, a simple model that introduces a small set of learnable context tokens at the input of a frozen CLIP and optimizes with both CLIP's image-text alignment objective and GCD's visual clustering objective. On Stanford and Clevr-4 datasets, OAK achieves state-of-the-art in accuracy and concept discovery across multiple categorizations, including 87.4% novel accuracy on Stanford Mood, surpassing CLIP and GCD by over 50%. Moreover, OAK produces interpretable saliency maps, focusing on hands for Action, faces for Mood, and backgrounds for Location, promoting transparency and trust while enabling adaptive and generalizable categorization.

LAVQA: A Latency-Aware Visual Question Answering Framework for Shared Autonomy in Self-Driving Vehicles

When uncertainty is high, self-driving vehicles may halt for safety and benefit from the access to remote human operators who can provide high-level guidance. This paradigm, known as {shared autonomy}, enables autonomous vehicle and remote human operators to jointly formulate appropriate responses. To address critical decision timing with variable latency due to wireless network delays and human response time, we present LAVQA, a latency-aware shared autonomy framework that integrates Visual Question Answering (VQA) and spatiotemporal risk visualization. LAVQA augments visual queries with Latency-Induced COllision Map (LICOM), a dynamically evolving map that represents both temporal latency and spatial uncertainty. It enables remote operator to observe as the vehicle safety regions vary over time in the presence of dynamic obstacles and delayed responses. Closed-loop simulations in CARLA, the de-facto standard for autonomous vehicle simulator, suggest that that LAVQA can reduce collision rates by over 8x compared to latency-agnostic baselines.

3DGEER: Exact and Efficient Volumetric Rendering with 3D Gaussians

3D Gaussian Splatting (3DGS) marks a significant milestone in balancing the quality and efficiency of differentiable rendering. However, its high efficiency stems from an approximation of projecting 3D Gaussians onto the image plane as 2D Gaussians, which inherently limits rendering quality--particularly under large Field-of-View (FoV) camera inputs. While several recent works have extended 3DGS to mitigate these approximation errors, none have successfully achieved both exactness and high efficiency simultaneously. In this work, we introduce 3DGEER, an Exact and Efficient Volumetric Gaussian Rendering method. Starting from first principles, we derive a closed-form expression for the density integral along a ray traversing a 3D Gaussian distribution. This formulation enables precise forward rendering with arbitrary camera models and supports gradient-based optimization of 3D Gaussian parameters. To ensure both exactness and real-time performance, we propose an efficient method for computing a tight Particle Bounding Frustum (PBF) for each 3D Gaussian, enabling accurate and efficient ray-Gaussian association. We also introduce a novel Bipolar Equiangular Projection (BEAP) representation to accelerate ray association under generic camera models. BEAP further provides a more uniform ray sampling strategy to apply supervision, which empirically improves reconstruction quality. Experiments on multiple pinhole and fisheye datasets show that our method consistently outperforms prior methods, establishing a new state-of-the-art in real-time neural rendering.

DINO-R1: Incentivizing Reasoning Capability in Vision Foundation Models

The recent explosive interest in the reasoning capabilities of large language models, such as DeepSeek-R1, has demonstrated remarkable success through reinforcement learning-based fine-tuning frameworks, exemplified by methods like Group Relative Policy Optimization (GRPO). However, such reasoning abilities remain underexplored and notably absent in vision foundation models, including representation models like the DINO series. In this work, we propose \textbf{DINO-R1}, the first such attempt to incentivize visual in-context reasoning capabilities of vision foundation models using reinforcement learning. Specifically, DINO-R1 introduces \textbf{Group Relative Query Optimization (GRQO)}, a novel reinforcement-style training strategy explicitly designed for query-based representation models, which computes query-level rewards based on group-normalized alignment quality. We also apply KL-regularization to stabilize the objectness distribution to reduce the training instability. This joint optimization enables dense and expressive supervision across queries while mitigating overfitting and distributional drift. Building upon Grounding-DINO, we train a series of DINO-R1 family models that integrate a visual prompt encoder and a visual-guided query selection mechanism. Extensive experiments on COCO, LVIS, and ODinW demonstrate that DINO-R1 significantly outperforms supervised fine-tuning baselines, achieving strong generalization in both open-vocabulary and closed-set visual prompting scenarios.