Geometric structure is essential for robotic manipulation and understanding the 3D world. Most VLA models rely on 2D representations of current observations, while geometry-aware approaches often require additional 3D information. However, geometry inferred independently from individual observations is often discarded rather than accumulated into a persistent task representation. Our key insight is that action generation needs a compact, task-relevant geometric abstraction that can be continuously updated, rather than a complete, renderable 3D reconstruction at every step. We introduce GeoMem-VLA, a VLA framework that internalizes explicit 3D supervision into dedicated geometric tokens and maintains them as temporal memory for action generation, without directly aligning the VLM's native visual embeddings with the teacher feature space. During training, a frozen geometry foundation model supervises learnable GeoPlan tokens through a geometric flow-matching objective over the current scene point map and future end-effector tracks. These tokens are integrated with the memory of observation representations, where historical information is retrieved, adaptively fused with current observations, and consolidated to retain event frames as the task progresses. At inference, the policy retrieves historical representations and fuses them with the current working representation to predict actions, without invoking the 3D teacher or explicitly reconstructing the scene. After benchmark-specific downstream adaptation, GeoMem-VLA achieves 98.9% average success on LIBERO and 91.10% / 91.24% on RoboTwin's clean/randomized settings. Across five real-world tasks, GeoMem-VLA achieves 78.0% average success. These results suggest that internalizing and remembering 3D scene understanding is a promising approach to robot control.
https://arxiv.org/abs/2610.29866
Autonomous mobile robots operating in dense crowds and cluttered sites require $360^\circ$ perception to react to surrounding hazards. While panoramic LiDAR offers full-surround coverage, learning representations within reinforcement learning (RL) is sample-inefficient under sparse rewards, and existing 3D pretrained point-cloud encoders are too memory-heavy for massively parallel RL simulation. We present RangeNav, a compact (183k parameters) LiDAR range-image encoder pretrained entirely on synthetic data. RangeNav uses anisotropic convolutions and circular padding to handle the extreme aspect ratio and azimuthal seam of range images. We pretrain the network using a navigation-aligned objective combining range reconstruction, semantic segmentation, and bird's-eye-view (BEV) prediction, then freeze it for policy learning. In simulation, adding semantic and BEV supervision increases success rates over a reconstruction-only baseline by an average of 32.03 and 23.73 percentage points (pp) in procedurally generated training and out-of-distribution realistic test environments, respectively. Zero-shot real-world deployments on a Unitree B2-W quadruped demonstrate robust navigation through dense crowds, low-light settings, and construction zones where limited-field-of-view depth policies struggle. Code and models will be open-sourced.
https://arxiv.org/abs/2610.29845
With the advent of foundation models, robots are increasingly capable of understanding and executing explicitly specified human tasks. Extending these capabilities to open-ended tasks, however, requires robots to infer appropriate actions from the observed scene. Household organization exemplifies this challenge, as robots must often act without detailed instructions, predefined object vocabularies, or known object-receptacle associations. While robots may follow instructions or learn organization preferences from demonstrations, inferring the latent organization of an occasionally changing scene and placing a novel object consistently with it remains challenging. We propose an open-world multi-attribute reasoning pipeline that grounds unknown objects across views and constructs an instance-level 3D scene graph for placement reasoning. First, a vision-language model proposes labels together with attributes such as content and packaging type, which are grounded through segmentation, depth, and multi-view association. Then, an LLM reasons over this representation to infer the latent shelf organization and rank scene-consistent object placements without explicitly user-provided preferences. The resulting placement rankings guide subsequent geometric and kinematic feasibility checks for execution. Our experimental evaluation shows that our approach leads to improved instance mapping and placement coherence over single-view language-grounding and geometry-only baselines, enabling autonomous placement of previously unseen objects without predefined object-receptacle mappings or class-specific retraining.
https://arxiv.org/abs/2610.29761
We present SpatialDex, a framework that learns 3D-grounded bimanual dexterity from RGB-only human videos through reconstruction-augmented visuomotor pre-training. The central idea is to learn human behavior together with its spatial context. From each video, we reconstruct the scene geometry and estimate the hand motion, align them at a shared video-level scale using the hand-size prior of 3D hand estimates. We retain the reconstructed geometry as both a policy input and a prediction target. The policy grounds the token grid of a pre-trained vision-language model in this geometry through gated residual fusion, is trained to predict future scene geometry alongside wrist and finger actions, and generates finger motion conditioned on the predicted trajectories of both wrists. Pre-trained on 1M human-video clips and fine-tuned with a small number of robot demonstrations, SpatialDex outperforms the state-of-the-art methods by 16.5 points on nine simulated dexterous tasks across two embodiments and by 38.3 points on six real-world bimanual dexterous tasks. Under a matched comparison, pre-training with reconstructed geometry improves success by 14.2 points over pre-training on hand motion alone.
https://arxiv.org/abs/2610.29716
Large-scale aerial resources provide a foundation for learning navigation across diverse environments. However, existing resources remain limited in providing (1) densely sampled flight demonstrations for learning different policy formulations and (2) broad environment coverage for standardized closed-loop evaluation. To address these limitations, we introduce AirPlayground, a unified benchmark for training and evaluating aerial navigation policies across diverse environments. We first construct 94 photorealistic 3D Gaussian Splatting (3DGS) environments spanning indoor, urban outdoor, and vegetation settings, with aligned collision geometry. Building on these environments, we develop a scene-adaptive pipeline for automated expert flight collection, yielding 94K demonstrations and 46.5M synchronized observation-state-action samples at 50 Hz. We further provide an extensible API that supports training different policy formulations on this shared corpus and evaluating them under unified protocols in the corresponding environments. We benchmark seven baselines spanning four policy paradigms across three navigation settings in seen and unseen environments, and further examine scaling behavior with ACT. Scaling ACT's training corpus from 5 to 70 scenes raises median success rates on seen and unseen environments from 30.6% and 16.8% to 41.2% and 29.4%, respectively. Through this shared training and evaluation framework, AirPlayground supports analysis of how offline prediction errors relate to closed-loop outcomes. Our analyses highlight the roles of prediction errors, execution protocols, and scene interactions in closed-loop performance.
https://arxiv.org/abs/2610.29699
Vision-language models (VLMs) achieve remarkable performance on 2D medical images, such as X-rays or pathology images, while 3D VLMs that directly process CT/MRI volumes lag behind due to limited 3D training data and the predominance of 2D foundation architectures. To leverage powerful 2D VLMs for analyzing medical volumes, which contain hundreds of slices, it is therefore necessary to focus only on those few slices, which are diagnostically relevant. To enable such a prioritization, we propose XGLASS, a method that transfers medical expertise from 2D X-rays to 3D CT for task-aware slice prioritization. XGLASS uses disease-conditioned importance maps derived from language-aligned radiologist gaze on X-rays and spatially aligns them with coronal CT projections, bringing gaze priors and CT content into a shared 2D representation. The fused gaze-prior/CT projection is then used to predict a disease-specific ordering of axial slices, prioritizing diagnostically relevant slices. XGLASS significantly improves budgeted retrieval over 2D and 3D-native baselines, achieving a top-10 hit rate up to 75.9% for the best-performing disease. In practice, this means that after inspecting only the top 10 slices (4.17% of a 240-slice scan), XGLASS identifies a diagnostically relevant slice in more than 75% of cases. For downstream slice-wise finding classification, an XGLASS-selected 30-slice candidate set reduces model calls by 87.5% while improving F1 by 11.3 percentage points over all-slice inference, 4.2 points over the strongest 2D selector, and remains competitive with 3D-native selectors while requiring up to 130.8 times less time and 10.9 times less GPU memory per finding, supporting its use as a lightweight preselector for volumetric CT analysis. The code and pretrained checkpoint are available on this https URL.
https://arxiv.org/abs/2610.26255
Text-to-motion generation and motion-to-text captioning are two fundamental tasks in human motion modeling, both grounded in the same underlying motion-text correspondence. Existing unified approaches mostly rely on autoregressive modeling, which imposes a fixed generation order and is therefore poorly suited to the bidirectional dependencies between language and motion, allowing early prediction errors to persist as fixed context and degrade both temporal coherence and cross-modal consistency. Masked discrete diffusion, which models sequences through iterative bidirectional prediction, allows prediction from bidirectional context without a fixed generation order. We therefore propose BiMoGen (Bidirectional Motion-text Generation), a unified masked discrete diffusion framework for bidirectional motion-text modeling. To stabilize training, we design Decoupled Uni- and Cross-Modal Training, in which masked pretraining first establishes cross-modal correspondence on paired motion-text sequences, after which supervised fine-tuning specializes the model for bidirectional generation. Masked diffusion nonetheless faces a train-test mismatch, as the model is trained on clean ground-truth context yet encounters self-generated and potentially erroneous context at inference, with errors committed under heavily masked states propagating through subsequent steps. We further introduce Generation-Aware Self-Correction that exposes the model to its own predictions during training and uses correction passes at early sampling steps for further refinement of previously unmasked tokens. Extensive experiments on HumanML3D and KIT-ML demonstrate competitive performance on both tasks, validating the effectiveness of the proposed two-stage training and self-correction designs. The project page is available at this https URL.
https://arxiv.org/abs/2609.35407
High-fidelity 4D physical simulation requires accurate physical parameter assignments that align with diverse material properties. Previous methods struggle with critical bottlenecks: they either rely on computationally intensive per-scene optimization or predict parameters from purely visual inputs. The latter suffers from inherent ambiguity, e.g., the inability to distinguish the stiffness of rubber from its appearance alone. To bridge this gap, we propose AutoPhyX, a text-conditioned framework for predicting spatially-varying physical properties. Since training robust feed-forward models requires fine-grained supervision, we first introduce a novel part-controllable data generation pipeline that decomposes complex 3D assets into semantically distinct components and pairs them with physically plausible, diverse parameters and corresponding text descriptions. Leveraging this dataset, AutoPhyX employs a cross-modal modulation mechanism where text features dynamically modulate visual features for precise physical grounding. By formulating property prediction within a voxel field, AutoPhyX ensures compatibility with diverse 3D representations, including meshes, point clouds, Gaussian Splatting, and NeRF. Experiments demonstrate that our method enables physically plausible, text-driven parameter assignment in a single forward pass. It achieves high accuracy and robust generalization to in-the-wild objects, paving the way for future physics-intensive embodied and robotic applications.
https://arxiv.org/abs/2610.29517
LEGO walkers are walking automata constructed from modular LEGO Technic bricks. They exhibit intriguing locomotion and are also highly reconfigurable through the modular Technic bricks, making them widely applicable in education, recreation, and robotics research. However, existing LEGO walkers are designed mostly by reusing existing single-actuator walking mechanisms. Although the locomotion is stable, their fixed walking gaits limit their adaptability and versatility. This paper presents the first computational solution capable of designing LEGO walkers that not only walk stably, but also support a variety of walking gaits. Our idea is twofold. First, we control the foot motion trajectory of each leg mechanically via a specifically-designed gear-linkage mechanism, in which we incorporate a pair of 3D-printed non-circular gears to work with a LEGO Technic mechanism. Second, we drive each leg by an independent motor and synchronize the timings of multiple leg motions via a simple yet adaptable control algorithm, enabling stable and versatile locomotion. Importantly, we take this modeling as a foundation to co-design the morphology and control of the LEGO walkers to meet the target locomotion and appearance. Leveraging our approach, we created a variety of LEGO walkers, demonstrating various walking gaits and appearances, and fabricated some of them to show their locomotion capability in the real world.
https://arxiv.org/abs/2610.29505
Full 3D attention is costly in video Diffusion Transformers. Block-sparse attention reduces this cost by computing only selected attention blocks. However, existing methods either overlook the potential to optimize sparsity allocation or rely on heuristic rules or local metrics to guide it. In this work, we observe that sparsifying attention in different layers and at different denoising steps affects final-video fidelity to substantially different degrees. Yet the identities of the most sensitive layer-step pairs remain nearly unchanged across inputs. Building on these findings, we propose Allocating Layer-step Latency by Output Tests (ALLOT), a training-free sparsity allocation method. Treating each layer-step pair as a cell, ALLOT estimates each cell's contribution to final-video error through (i) cell probing and (ii) trajectory calibration, then jointly optimizes cell densities under a measured-latency budget. It preserves existing selection strategies and kernels, replacing only the allocation component. At equal or shorter wall-clock time, ALLOT outperforms every compared allocation strategy with XAttention, with an 11.8 dB PSNR gain on HunyuanVideo. Extensive experiments across four additional published pipelines on Wan2.1, Wan2.2, and HunyuanVideo show that ALLOT outperforms or ties the default in all nine settings.
https://arxiv.org/abs/2610.29441
Real-world physical sensing exhibits complex, heterogeneous noise patterns that deviate significantly from idealized simulation, posing a fundamental bottleneck for sim-to-real transfer. Existing sensor modelings typically treat depth noise as a monolithic black-box process, overlooking the distinct physical mechanisms that govern different error modalities. In this work, we introduce a physics-grounded paradigm that disentangles monolithic noise into two complementary modalities: \textbf{sensing invalidation} and \textbf{measurement inaccuracy}, enabling a tailored treatment of noise sources based on their physical origins. Building on this insight, we propose \textbf{PRISM}, a tripartite framework that distills 3D Visual Foundation Model features as rich spatial-semantic priors for physics-based reasoning. To address the inherent sparsity and class imbalance of invalidation regions, we develop Hierarchical Positive-Prioritized Supervision, integrating multi-scale positive-weighted objectives with a positive-preserving dynamic hard mining strategy to enforce precise artifact delineation. Extensive benchmarks demonstrate that PRISM achieves state-of-the-art fidelity in noisy depth synthesis. Furthermore, downstream robotic experiments show that PRISM facilitates a \textbf{92.1\%} average success rate in the real world, marking a significant improvement over monolithic baselines. Our project page is available at: \resizebox{\linewidth}{!}{\href{this https URL}{this https URL}.}
https://arxiv.org/abs/2610.29432
In implicit neural representations (INRs), 3D signals are encoded directly in network weights, making stable weight-space operations essential. However, permutation symmetry causes independently trained yet functionally equivalent models to appear distant in weight space, often inducing large loss barriers under linear interpolation. Existing neuron alignment methods either commit early to discrete matches and become trapped in inconsistent local optima or rely on computationally expensive gradient optimization over continuous relaxations, both of which are unreliable in challenging cases such as NeRF. We adopt progressive alignment, a layer-wise strategy that follows inter-layer dependencies from input to output. At each layer, a correlation matrix is estimated and aligned with a discrete anchor obtained by the Hungarian algorithm, after which temperature-controlled spectral annealing gradually sharpens the correlation structure toward a permutation matrix. This update is deterministic and lightweight, avoiding the inner-loop gradient optimization required by continuous relaxation methods. A final refinement stage further optimizes the network layer-by-layer via weight matching, starting from the aligned solution. Experiments on MLPs, CNNs, and Neural Radiance Fields (NeRFs) show that the proposed method reliably recovers neuron correspondences and achieves stable alignment across architectures. In practical experiments aligning independently trained NeRF models, our method reduces the loss barriers under weight interpolation compared with existing approaches.
https://arxiv.org/abs/2610.29396
3D semantic occupancy prediction estimates voxel-wise occupancy and semantic labels throughout a scene. Our experiments reveal that conventional sparse transposed-convolution upsampling yields poorly differentiated voxel responses and that successive decoder layers discard previously correct semantic predictions. To address these issues, we present Feature-Diverse Upsampling and Memory-Aware Decoding for 3D Occupancy Prediction (FM-Occ), comprising two components. First, 3D Sparse Subvoxel Shuffle (3S) mitigates feature homogenization by constructing candidate child features in the channel dimension, jointly transforming and normalizing them, and then rearranging them onto a finer sparse grid with explicit feature-coordinate correspondence. Second, Entropy-Regulated Layer-wise Memory (ER-LM) retains semantic information across decoder layers by maintaining a layer-wise memory for each query through gated recurrent updates and using prediction entropy to regulate the historical information propagated to the next layer. During training, a class-weighted entropy loss Lcwe provides label-guided regularization for matched semantic queries. On SemanticKITTI, 3S increases the spatial coefficient of variation of normalized voxel-wise root-mean-square (RMS) responses from 0.41% to 4.03%, while ER-LM reduces semantic degradation and improves error correction across decoder layers. The combined FM-Occ improves mIoU from 12.16% to 12.73% with negligible memory overhead. ER-LM further improves mIoU in the BEV-based DISC framework from 16.78% to 17.17%, supporting its applicability across 3D voxel-based and BEV-based architectures.
https://arxiv.org/abs/2610.29388
Robots deployed in human-centered environments need maps that support both geometric navigation and open-vocabulary semantic queries beyond a fixed label set. A key challenge is enriching such maps with semantic information without relying on manually annotated segmentation masks or closed-set semantic categories. Existing geometry-driven active mapping maximizes coverage but treats semantics as a passive by-product of the collected trajectory, whereas active semantic mapping typically depends on supervised semantic segmentation. We present AOV-GS, an active 3D Gaussian mapping system that guides exploration using frozen foundation models. During coarse exploration, AOV-GS re-ranks geometrically promising candidate views according to CLIP mask-embedding novelty computed from SAM instance segmentation, biasing exploration toward semantically informative regions. Frozen SAM and CLIP provide instance masks and language-aligned embeddings throughout mapping, while a lightweight scene-specific language field is fitted offline directly to the collected CLIP embeddings rather than manually annotated masks, enabling open-vocabulary text-query segmentation on novel views. In a controlled comparison using the same SAM-CLIP perception stack, AOV-GS improves average language-field mIoU over geometry-only active planning by 1.3 percentage points on Replica (34.6% on eight scenes) and by 1.8 points on Matterport3D (23.1% on five scenes). With CorrCLIP, AOV-GS reaches 35.6% on Replica, 2.0 points above the matched Geometry-based + CorrCLIP baseline. Novel-view PSNR remains comparable to the geometry-only baseline on both datasets while depth L1 improves on Replica. Code is available at this https URL.
https://arxiv.org/abs/2610.29374
Spatial reasoning requires more than access to geometric tools: a reasoner must decide when direct observation is sufficient and when to reconstruct, measure, or change reference frames. We present SpatialBrain, a framework that retains experience about these decisions while keeping its vision-language backbone fixed. The key distinction is between evidence about a scene and experience about how to reason with that evidence. 3DFusionMemory fuses current-scene observations into queryable spatial relations. A Spatial Skill Library provides reusable tool procedures, and Spatial Cognition associates task conditions with experience about when and how to use memory and skills. Completed trajectories are abstracted into process records and consolidated between task blocks, allowing later decisions to draw on earlier execution without transferring scene coordinates or reference answers. Across five spatial-reasoning benchmarks, SpatialBrain improves on direct answers from the same backbone on four and ties on the fifth. Component comparisons, experience-coverage curves, and traced resource decisions characterize how the framework accumulates and uses these resources.
https://arxiv.org/abs/2610.29317
World Action Models (WAMs) have shown strong potential for robot manipulation by jointly learning visual dynamics and executable actions. However, their video and action objectives do not explicitly constrain the 3D geometry that is crucial for precise manipulation. Recently, geometry-aware WAMs introduce depth or 4D prediction branches, whereas representation alignment methods supervise intermediate WAM features with pretrained 3D foundation models. Yet we find that such pointwise alignment benefits the video features but degrades action learning, since dense geometric feature targets overly restrict compact action features. To address this limitation, we propose Relational Geometry Forcing (RGF), which transfers geometric priors from a frozen 3D foundation model by aligning pairwise feature relations instead of pointwise features. Specifically, RGF applies three complementary relational objectives, where intra-frame and cross-view relations transfer scene geometry and multi-view correspondence to the video features, while action-temporal relations couple action dynamics to geometric scene evolution. Applied only during training, RGF requires no explicit geometry branch or extra inference module. Extensive experiments in simulation and the real world demonstrate the effectiveness of Relational Geometry Forcing, showing consistent gains across diverse manipulation scenarios and WAM paradigms.
https://arxiv.org/abs/2610.29282
As-built BIM reconstruction from point clouds demands semantic, volumetric, parametric, and editable building components, beyond 2D layouts or planar abstractions. Yet existing scan-to-BIM pipelines often rely on handcrafted rules, heuristic optimization, or manual correction, while learning-based methods fall short of recovering wall thickness, opening geometry, and component-level BIM structure. We introduce CRAB, a BIM-ready cuboid reconstruction framework that encodes walls, doors, windows, beams, and columns as compact architectural cuboids with IFC-compatible semantics and parameters. CRAB bridges point clouds and editable BIM by combining ProposalNet, which predicts coarse component cuboids, with AssemblyNet, a relation-aware Flow Matching model that refines geometry, predicts structural relationships, and enables automatic structural correction and human-guided editing. To address limited cuboid-level supervision, we extend Structured3D with component-wise cuboid annotations and construct cuboid-based procedural scenes. Experiments show CRAB surpasses floor plan and plane-based alternatives in component completeness, geometric accuracy, and generalization. For interior wall thickness estimation, CRAB achieves a mean error of 2.80 cm, with 79.50% of matched walls within 3 cm. The project page is available at this https URL.
https://arxiv.org/abs/2610.29257
Predictive representations for robot manipulation should connect scene geometry with the state changes induced by actions. Our intuition is that understanding what an action will change starts with understanding where things are: complementary RGB and depth observations ground a common scene representation from which to infer those changes. Guided by this intuition, we introduce \textbf{DepState-JEPA}, a dual-stream RGB--Depth joint-embedding predictive architecture that connects spatial understanding with action-conditioned prediction through a shared latent state, \textbf{DepState}. Built on V-JEPA~2.1, the framework uses depth both as a complementary observation stream and to ground both modalities in metric 3D space, aggregating shared information into DepState while retaining modality-private features for decoding. Ray-conditioned readouts predict local features without target depth; after an action-conditioned transition, the same decoders predict future RGB and depth features using the predicted state and fixed initial private features. Robot-state readouts provide recursive proprioceptive inputs; object-trajectory supervision and calibrated forward kinematics shape state learning through motion targets and joint--TCP consistency. Comparisons show stronger RGB-encoder transfer after RGB-D adaptation, feature completion, conditioned visual forecasting, and manipulation success. Within-model ablations identify contributions from depth content, metric coordinates, and future-state matching, with additional control gains from FK regularization.
https://arxiv.org/abs/2610.29246
Vision-Language Models (VLMs) excel at 2D recognition but often fail to maintain consistent 3D relations across sparse viewpoints, leading to spatial hallucinations under occlusion and perspective changes. Existing methods commonly fuse all frames with dense geometry, incurring substantial computation and context overhead while encouraging blind reliance on imperfect estimators. We present Think-Capture, a query-conditioned framework that first identifies task-critical objects and then acquires their geometry on demand. A 3D toolkit lifts zero-shot masks into a common coordinate frame, and a Symbolic 3D State Abstraction compresses object and camera states into structured text. Hierarchical Group Relative Policy Optimization (GRPO) trains the VLM on cases in which geometry either helps or misleads, teaching it when to use tool evidence and when to fall back on visual cues. Think-Capture-4B reaches 54.4% on MindCube-Tiny (+21.5 points over its backbone), 39.1% on the MMSI-Bench positional subset (+7.9), and 42.8% on ViewSpatial (+5.2). It also remains robust under coordinate and mask perturbations. The full stack fits one 24 GB GPU and serves as an asynchronous high-level reasoner on an Agibot G1. Our code is available at this https URL.
https://arxiv.org/abs/2610.29232
Recovering dynamic 3D scene geometry, appearance, and the physical motion states that govern scene evolution from multi-view videos is important yet challenging. A key difficulty is that appearance can be deceptive. Under rendering supervision alone, a model may fit the observed images by changing color or opacity to compensate for inaccurate deformation, rather than recovering the true evolving geometry and motion. This issue is especially severe in scenes with fast or locally complex motion, where inaccurate trajectories can still produce plausible images over the observed frames, but lead to unreliable reconstruction and discontinuous future prediction. In this paper, we propose PhyMo, a framework for learning evolving dynamics from multi-view videos. We introduce a Motion-Aware Trajectory Representation that drives the model to recover accurate motion from deformation, rather than relying on appearance shortcuts to explain observed motion. On top of this, we propose a Continuous Motion Evolution module with bidirectional kinematic continuity constraint to promote temporally consistent evolution of displacement, velocity, and acceleration. Together, these designs yield geometrically faithful interpolation and continuous motion extrapolation. Experiments on four dynamic scene benchmarks demonstrate state-of-the-art performance.
https://arxiv.org/abs/2610.29107