Visuomotor Policy Learning

Generative visual models offer a foundation for learning representations of physical dynamics, yet their extension to continuous control raises a fundamental question: do visual prediction and action generation require separate computational pathways? Existing approaches usually introduce trainable action heads or separate action experts to bridge low-dimensional states and high-dimensional visual representations. In this work, we explore whether the visual backbone's existing capacity can also support control when actions are expressed in a compatible representation. Thus, we introduce PatchWAM (Patch World-Action Model), which treats continuous actions as another type of patch through a fixed mapping called Action-as-Patch. This allows a single model to predict both how the robot should move and what the scene may look like afterward. Visual prediction and action generation become parts of the same generative process, without a dedicated action head or separate action expert. Experiments with subsampled training windows show gains over a matched dual-expert control, while benchmark evaluations reach 91.8% success rate on LIBERO-Plus and 96.12% on RoboTwin 2.0 in a full-data setting with additional augmented demonstrations. More broadly, the result suggests that capability need not be added where it can be inherited: the constraint on extending a generative backbone is the interface a new signal is written in, not the capacity to model it.

Vision-language-action policies benefit from geometric supervision, but current-frame geometry alone does not explicitly describe the changes associated with manipulation. This design is motivated by the goal of learning an embodiment-agnostic visual interface that can be pretrained across robot and egocentric video before robot-specific action alignment. We introduce Geometry-Change VLA (GC-VLA), which learns to predict multiview future-current geometry-change tokens from current observations. Offline frame pairs define a nominal 0.5-second prediction horizon; future observations are used only to construct training targets. Stage 1 trains a geometry-change vision-language model (GC-VLM). Stage 2 introduces a continuous ActionExpert and aligns it with robot actions while stopping action-flow gradients at the VLM interface. Stage 3 enables these gradients to update the trainable VLM components jointly with the ActionExpert. Stage 4 freezes GC-VLA and applies Geometry-Conditioned Residual Flow (GCRF), using a binary intervention router and a single bounded residual velocity policy learned from closed-loop feedback. GC-VLA achieves 95.20% success on LIBERO, and GC-VLA with GCRF achieves 99.55%. Inference uses current observations and the learned GC representation without executing the offline target encoders.

Robot manipulation policies must generalize across visual shifts while preserving scene context relevant to action. General-purpose vision encoders are not tailored to visuomotor control, while object-centric approaches often use segmentation masks as hard filters that discard potentially useful context. We propose RoboMP-DINOv2 (Robotics Mask-Prompted DINOv2), a full-scene vision encoder that treats masks as spatial prompts rather than visibility filters. It extracts dense DINOv2 features from the full observation, injects learned region-specific embeddings at masked locations, and jointly contextualizes prompted and unprompted tokens for action prediction. We further introduce masked-region color randomization (MCR) to improve appearance robustness, yielding RoboMP-DINOv2-MCR. Across seven simulated manipulation settings, RoboMP-DINOv2 achieves 60.7% success under spatial shifts and 59.7% under scene clutter, compared with 50.7% and 41.0% for a DINOv2-based Diffusion Policy. Under unseen object colors, RoboMP-DINOv2-MCR achieves 72.5% success versus 35.1% for the strongest color-randomized baseline. Additional experiments and representation analyses show improved robustness while preserving behaviorally relevant scene information. Code is available at https://github.com/han20192019/RoboMP_DINOv2.

Dormant tree pruning is labor-intensive yet essential for maintaining modern high-productivity fruit orchards. In this work, we focus on pruning of modern planar tree training systems - V-Trellis apples and UFO cherries - where trunks and primary branches are trained into approximately planar walls. We introduce an end-to-end pipeline to learn a closed-loop visuomotor controller for robotic pruning. This controller is trained entirely using simulation and synthetically generated data and deployed in real orchards in a zero-shot manner. The pipeline comprises synthetic generation of planar orchard tree meshes, construction of a physics-based orchard simulator, automated collection of successful pruning trajectories via motion planning, and policy learning with a novel hybrid reinforcement-learning algorithm that combines offline demonstrations with online simulated rollouts. The controller uses optical-flow inputs from a wrist-mounted camera - avoiding the need for full 3D-reconstruction - and continuously guides the cutter through cluttered branch environments to a specified cutpoint with correct tool orientation. In exhaustive simulated task-space evaluations over 3,000 pruning points, the policy attains 49.9% success on V-Trellis apples and 46.0% on UFO cherries. We validate the learned controller across 38 physical trials - comprising 28 outdoor field trials in commercial and experimental orchards and 10 indoor laboratory tests - demonstrating zero-shot sim-to-real transfer. The learned policy also outperforms a classical RRT-Connect baseline on physical hardware in laboratory trials.

Adapting visuomotor policies to new manipulation tasks often requires substantial manual engineering or teleoperated data collection. Simulation can provide task-specific data at scale, but constructing the scene, designing expert behavior, and configuring data generation still require significant per-task effort. We present ARSTAG, an agentic Real2Sim2Real system that turns a single RGB image and a natural-language instruction directly into robot policy-learning data. A hierarchy of language agents constructs a task-scoped simulation scene, generates robot-feasible demonstrations, and expands the training distribution through task-consistent randomization, while a coordinator agent manages cross-stage feedback and recovery. Across seven manipulation tasks spanning grasping, placement, and stacking, the ARSTAG-generated demonstrations enable sim-to-real transfer of three visuomotor policy architectures to a dual-arm robot, with pi0.5 achieving an average real-world success rate of 74.6%. Ablations show that task-consistent randomization substantially improves robustness, and policy performance increases with generated dataset size. Project webpage: https://boweili666.github.io/ARSTAG/.

Robot visuomotor policies are commonly formulated as autoregressive, diffusion-based, or more recently, flow matching models. Among them, Action-to-Action (A2A) flow matching improves inference efficiency by initializing generation from historical action priors rather than stochastic noise. However, stale historical motion patterns and entangled global visual representations can jointly reduce robustness under spatial out-of-distribution (OOD) shifts and visual distractors. In this work, we propose SlotFlow, an object-centric flow matching policy for robust visuomotor manipulation. SlotFlow decouples scene observations into semantic ("what") features and lightweight image-plane spatial ("where") cues to provide object-aware policy conditioning and current-state grounding. The semantic representation suppresses irrelevant background correlations, while the spatial cue improves adaptation to shifted object configurations. Extensive simulation and real-world experiments demonstrate improved robustness under visual distractors and severe spatial perturbations while preserving the low-step inference efficiency of A2A. Controlled initialization and perception ablations further identify object-centric grounding as a major source of the gains and show that it complements, rather than replaces, useful historical motion priors.

Learning visual policies for locomotion and manipulation requires coordinating contact with the environment and can incur substantial computation and GPU memory costs. First-order policy gradients (FoPG) reduce training cost through differentiable simulation, but local optimization can converge to unintended contact patterns. To address this shortfall, we propose Sampling-Guided Policy Search (SGPS), which couples recurring action-target refinement by sampling-based model-predictive control with first-order policy optimization. Behavior cloning initializes the policy from sampled actions; training then alternates sampling-based refinement with short-horizon FoPG updates under perturbed initial states and randomized dynamics. For visual policy training, we use a decoupled FoPG formulation that excludes rendering from the computation graph, enabling direct learning from depth observations without a state-policy teacher. On a single GPU, SGPS learns policies for locomotion, obstacle traversal, crate pushing, and bimanual carrying on simulated Unitree Go2 and G1 robots. Our experiments further show that refinement improves policy learning beyond initialization and tracking alone. For hardware deployment, the distilled policy transfers zero-shot to a real Go2 and uses onboard depth to autonomously trot, crawl, clear hurdles, and switch between these behaviors.

Visuomotor policies for multi-fingered dexterous manipulation are highly sensitive to camera viewpoint shifts. To achieve view invariance, recent methods increasingly rely on explicit 3D modalities like RGB-D or point clouds, which can introduce hardware dependencies, calibration requirements, and vulnerability to sensor noise during real-world deployment. In this work, we show that view-invariant control can be achieved without explicit test-time 3D sensing by encoding geometric knowledge into the visual representation during simulation. We present AnyViewDex, an asymmetric training pipeline that combines multi-view contrastive alignment with privileged 3D geometric supervision. By regressing absolute 3D object coordinates during simulated training, this auxiliary objective provides a geometric grounding signal that mitigates the spatial collapse of the globally pooled contrastive embedding. At deployment, the policy operates zero-shot using only uncalibrated monocular RGB and proprioception. We validate this approach across both reinforcement learning and student-teacher distillation. In hardware evaluation on an xArm7 with a 16-DoF LEAP Hand, AnyViewDex reaches 76.7% grasping success across eight unseen objects and six uncalibrated viewpoints (480 trials; 2,400 across all ablation conditions), indicating that geometrically grounded monocular policies transfer zero-shot without test-time depth. Project Page: https://anyviewdex.github.io/

Imitation-learned robot policies are frequently overfit to the visual conditions present in their training demonstrations. Consequently, variations in object color or background appearance often induce substantial performance degradation. A common mitigation strategy is to acquire additional demonstrations in each novel visual context; however, this approach is resource-intensive, requiring repeated access to a robot, a controlled environment, and human operation for every appearance condition to be covered. We introduce ReShoot, a framework that synthesizes visual diversity by re-rendering previously recorded demonstrations under altered appearances, thereby shifting the burden from data collection to generation. A vision-language model captions the scene, edits a targeted attribute (e.g., background, object color, or material), and an edge-conditioned video generator re-renders both camera views to match. The instruction is updated accordingly. The action sequence and proprioceptive trajectory are copied verbatim without relabeling, so each generated episode retains the recorded action and proprioceptive labels. On LIBERO, a policy trained on an equal mixture of recorded and re-rendered demonstrations matches the performance of recorded-only training (96.5% vs. 96.9%). Moreover, the mixed training set improves robustness to scene perturbations on LIBERO-Plus (85.5% vs. 82.3%). Across two physical robotic platforms, deploying ReShoot with 43 and 100 pre-collected demonstrations increased the success rate on recolored objects from 0.0% to 42.9% and 47.5%, respectively, while maintaining performance under the original recorded appearance.

Humanoid loco-manipulation requires adaptive whole-body coordination to seamlessly integrate locomotion and physical interaction. Despite recent advances, learning autonomous loco-manipulation remains challenging due to the scarcity of diverse, physically executable robot-object interaction data and the difficulty of learning unified whole-body control directly from onboard observations. We present ViLoMan, a scalable framework for autonomous humanoid loco-manipulation. ViLoMan first transforms partial kinematic demonstrations of human-object interactions into complete, physically executable robot trajectories. It then leverages these trajectories within a teacher-student distillation framework to learn a unified policy that maps egocentric depth observations and proprioceptive measurements directly to joint-level whole-body actions. During deployment, the policy requires neither reference motions nor intermediate commands. We evaluate ViLoMan on door-closing tasks across diverse door configurations and robot initial conditions in both simulation and the real world. Experimental results demonstrate that a single policy enables a Unitree G1 humanoid to complete the full task using only onboard depth sensing and proprioception, while generalizing robustly across task variations and transferring effectively from simulation to reality. Project page: viloman-anonymous.pages.dev.

Vision-Language-Action (VLA) policies commonly run Vision-Language Model (VLM) backbones with billions of parameters at every policy inference, which costs latency and energy. We revisit a decoupled alternative for multi-task manipulation: separate vision and language encoders whose representations condition a compact action head. We run a standardized comparison that varies the vision encoder, the language encoder, and the action head while holding the demonstrations, the training-step budget, the tasks, the evaluation protocol, and the measurement platform fixed, against seven VLA baselines. The resulting Decoupled Embodiment Model (DEM) combines a fine-tuned DINOv3 vision encoder, a frozen NeoBERT language encoder, and a MeanFlow head that generates an action chunk in one forward pass. On 18 RoboCasa tasks evaluated with held-out instruction paraphrases and randomized scenes, DEM reaches 55.6% mean success against 56.9% for GR00T N1.7 and 54.6% for $π_{0.5}$, and on three real-robot tasks it reaches 66.0% against 68.0% for GR00T N1.7. On the same workstation, DEM needs 6.1,ms per policy forward pass, a maximum throughput of 162.7 policy calls per second, and draws an estimated 2.07,J of GPU energy per call, eight to seventeen times the throughput and six to fifteen times less energy than these VLM-backbone policies. Within this trained-task regime, DEM sits on the observed success--latency--energy frontier and provides a strong, efficient baseline for language-conditioned robot skills.

Underwater visuomotor policy learning has focused primarily on single manipulators, while bimanual imitation learning has been studied largely in air. We present ULOHA, an underwater bimanual robot learning platform that combines custom-designed leader--follower hardware with software extensions to LeRobot, integrating teleoperation, multi-view sensing, demonstration collection, policy training, and autonomous deployment. Real-robot experiments demonstrate a range of coordinated underwater bimanual behaviors, including inter-arm transfer, shared-object manipulation, and buoyancy-driven interception. We evaluate ACT, Diffusion Policy, and the vision--language--action model SmolVLA on the platform. We investigate how learning methods and execution strategies developed for manipulation in air perform underwater, examining bubble disturbances, buoyancy-driven object motion, action-execution horizons, and real-time chunking. A separate single-arm study examines policy transfer between air and water and shows that demonstrations spanning both media support execution in both under the tested conditions. ULOHA provides a unified experimental platform for studying underwater bimanual robot learning under the coupled perceptual and physical effects of underwater environments. Additional material: https://mertcookimg.github.io/uloha/

Visuomotor policies observe both the task scene and the acting embodiment, allowing embodiment-specific visual cues to influence action prediction. We study this phenomenon as visual embodiment dependence (VED) and show, through cue-conflict interventions across representative policies, that visible robot configuration can become a shortcut to task progress. Rather than eliminating VED, we argue that it should be structured around embodiment information that supports control and generalization. We realize this through embodiment canonicalization in 3D point clouds, replacing the original embodiment with a canonical end-effector representation (CER) that preserves control-relevant geometry while abstracting embodiment-specific morphology. Its editable form further enables configuration-decorrelation augmentation for unfamiliar robot configurations. Experiments show that embodiment canonicalization substantially improves human-to-robot policy transfer without robot demonstrations, while simply removing the embodiment is insufficient without preserving control-relevant geometry. We further find that CER itself can become a configuration shortcut when robot configuration becomes decoupled from task progress; configuration-decorrelation augmentation mitigates this failure mode and restores robust recovery without sacrificing performance on seen configurations. Together, these results show that robust visuomotor learning benefits from structuring, rather than removing, visual embodiment information. Project website: https://tonyfang.net/ved

Human videos provide demonstrations of dexterous manipulation but lack robot-executable actions and tactile measurements. We present UniDex-ViTac, a framework that uses human-video-guided simulation to generate robot demonstrations paired with fingertip contact observations for training a deployable visuo-tactile policy. Object-specific residual reinforcement learning specialists adapt annotated human-object interaction references to a robotic arm-hand system. Their successful rollouts pair final robot action targets with robot-side fingertip contact observations. From 50 human demonstrations across ten objects, we collect 10,000 simulated trajectories to train a single Action Chunking with Transformers (ACT) based generalist. The policy combines point clouds, proprioception, and four binary contact signals encoded through fingertip labels and a separate token, without requiring human references or privileged object identity and pose at deployment. The contact-augmented configuration achieves 68.3% macro-average success in simulation, compared with 55.5% for the point-cloud-only baseline. Without real-robot demonstrations or policy fine-tuning, it succeeds in 73/110 physical trials (66.4%) across six seen and five unseen objects, compared with 60/110 (54.5%) for the baseline, an increase of 11.8 percentage points. These results support the feasibility of learning a unified visuo-tactile dexterous manipulation policy from video-guided simulated interactions. Project page: https://unidex-vitac.github.io/

Learning chunk-based visuomotor policies for long-horizon robot manipulation remains challenging. Recent action-chunking methods have shown promising performance by predicting temporally extended action sequences. However, their failures are often dominated by prediction errors at a small number of critical timesteps rather than uniformly poor predictions across the entire action chunk, making uniform refinement inefficient and insufficiently targeted. To address this bottleneck, we propose Uncertainty-Guided Refinement (UGR), a sparse refinement framework for chunk-based visuomotor policies. Specifically, UGR follows a coarse-to-refine design: it first predicts a full action chunk, estimates per-step temporal uncertainty from the coarse hidden states, and applies residual correction only to the most uncertain timesteps selected by a binary mask. The uncertainty branch is decoupled from the coarse action predictor, enabling clean attribution of the refinement gains to uncertainty-guided correction rather than additional predictor capacity. Extensive experiments on five dual-arm manipulation tasks from the RoboTwin benchmark show that UGR achieves the best success rate on four tasks, improves over the ACT baseline by up to 13% absolute, and outperforms both full-chunk and position-agnostic block refinement in ablation studies.

Recent advances in robot imitation learning have produced visuomotor policies that predict actions directly from visual observations. Yet visually similar scenes can require different actions as target position, object height, or contact geometry changes. Pretrained RGB features may map these geometrically distinct states to similar policy inputs, while simply adding depth requires the policy to learn RGB-depth correspondence from the same limited demonstrations used to learn control. We introduce StereoPatch, a patch-aligned RGB-depth representation that binds registered metric geometry directly to the RGB patches used for action prediction. On a shared 2-D patch grid, asymmetric cross-attention incorporates depth information into the corresponding RGB features before action decoding. The resulting StereoPatch Tokens provide a geometry-aware visual representation that can condition general visuomotor policies without changing their underlying learning objectives. Across six real-robot tasks, StereoPatch achieves higher closed-loop success than appearance-only, geometry-only, raw RGB-D, and late-fusion baselines. Additional experiments across three simulation suites evaluate compatibility across visuomotor policy architectures, spatial generalization, and operating limits. Results suggest that resolving control-relevant geometric ambiguity benefits from aligning depth directly with the visual features used for action prediction, rather than supplying it as an independent modality. Project page: https://aus.bot/research/stereopatch/.

Integrating contact information into visuomotor policies remains an open problem. Touch is essential to robust manipulation, yet most modern policies, including pretrained vision-language-action (VLA) models, operate from vision and proprioception alone. Existing approaches to closing this gap require specialized tactile hardware, add separate tactile encoders, or commit to non-image policy backbones, all incompatible with the modern paradigm of image-conditioned policies built on pretrained 2D visual representations. Our key insight is that the bottleneck is not the contact information itself, but how it is delivered: when contact signals are exposed in the same spatial frame as the scene the policy already attends to, they become directly usable by any image-conditioned policy without architectural changes. We operationalize this insight in Visible Touch, paired with a custom low-cost magnetic contact sensor that is open-sourced and fabricated from off-the-shelf parts via a parametric CAD-to-mold pipeline. Across the LIBERO benchmark, Visible Touch improves BC-Transformer success by 15.7 percentage points on average in the 2-view setting, with similar gains in the 1-view setting; controlled comparisons show that the contact-integration strategy strongly affects how effectively tactile information is used. The pattern holds when fine-tuning pretrained VLAs: miniVLA on LIBERO gains 25 percentage points on average, and $π_{0.5}$ on four real-world contact-rich tasks gains 30 percentage points with our custom sensor.

Recent advances in Vision-Language-Action (VLA) models have improved robotic manipulation, yet observation-to-action learning remains limited by a fundamental transition realizability gap, manifested in three tightly coupled problems: (i) Transition ambiguity. Visually similar current observations may correspond to different manipulation phases and imply different subsequent transitions. (ii) Prediction--execution mismatch. A visually plausible predicted future observation does not necessarily correspond to a physically realizable transition. (iii) Experience--realization mismatch. A historically executable action pattern may not necessarily realize the intended transition in the current scene and therefore requires context-aware adaptation. Accordingly, we propose UniMPA, a Unified Memory-Prediction-Action model that addresses these problems through a shared action-grounded transition interface. (i) UniMPA introduces Persistent-Selective Future Prediction to resolve transition ambiguity by modeling the intended future state evolution. A persistent latent stream continuously tracks task-level progress, while a transition-critical pixel stream selectively resolves fine-grained interaction changes through memory-grounded prediction. (ii) To assess the physical executability of the anticipated transition, the predicted transition queries a temporal Visual-Action Memory Bank. The bank retrieves historically realized visual-action experience, grounding future prediction in executable evidence. (iii) To adapt executable experience to the current scene, an Action-Visual Memory Bank retrieves visually grounded action prototypes from historical action evolution. Prototype-Biased Flow then shifts the flow source toward a historically supported action manifold for context-aware refinement.

Mobile manipulation requires perceptual evidence at different spatial scales for base motion and arm control, while the two action modalities remain kinematically coupled. Existing policies often employ specialized action generation for different subsystems but condition heterogeneous action branches on a shared perceptual representation, leaving subsystem-specific perception-action correspondence implicit. We present MoPA, a framework that aligns perceptual conditioning with mobility and manipulation while preserving coordination at the action level. Dual Perceptual Streams employ two mutually masked query banks to extract separate perceptual representations from a shared vision-language context. Perception2Action Adaptation jointly updates each query bank and its corresponding action stream at every layer of a structured Mixture-of-Transformers decoder, while enabling information exchange between the two action streams. Coupled conditional flow matching learns a joint vector field for coordinated generation of both action chunks. On the ManiSkill-HAB benchmark, MoPA achieves state-of-the-art performance across all three task suites. Across four real-world tasks, MoPA achieves a mean full-task success rate of 76.3%, outperforming the best baseline by 12.5 percentage points. Ablation studies and further analyses validate the effectiveness of the proposed design. Website is available at: https://mopa-policy.github.io/.

Standard behavior cloning supervises actions without explicitly constraining the future representation paired with each demonstrated action chunk. We introduce JEPA Policy, a diffusion-free framework that uses the action chunk and its observed future representation as paired training targets. Action and future-representation tokens interact in a shared Transformer and are refined through two forward passes. Future prediction can therefore shape the representation used to generate actions. Dual-branch and gradient-routing controls attribute the gain to this shared topology rather than to an auxiliary prediction head alone. Across nine simulated tasks, JEPA Policy improves mean success over the action-only MIP baseline and outperforms Diffusion Policy under the evaluated configurations, while adding 0.29 ms to MIP's model latency. A five-task, 630-episode physical-robot study produces the same pooled ranking. Further audits find no complete representation collapse under action supervision and identify a task-conditioned failure-ranking signal in future-prediction error. These results support paired future-representation supervision as a practical approach to low-latency visuomotor imitation without iterative generative sampling.

Focusing on spatially localized, control-relevant visual cues has been shown to improve data efficiency in visuomotor policies by reducing the need to model task-irrelevant visual variation. Existing methods often impose this focus through input preprocessing, such as cropping control- or object-centric regions in RGB images or point-clouds. However, it remains underexplored whether such localized features can be exposed directly from commonly used convolutional neural network (CNN) encoded features. In this paper, we show that intermediate CNN features preserve localized visual context for control, but existing pooling methods fail to aggregate it effectively. We introduce FocusPool, an attention pooling module that selectively aggregates intermediate visual features according to their relevance to the robot's current proprioceptive context. The resulting pooled representation captures task-progressive, control-relevant local information and is used directly for policy learning. Across simulation and real-world experiments, FocusPool improves policy success rates over pooling and explicit local focus methods by 36.2% and 41.2%, with training only 5.8% of encoder parameters.

Vision-Language-Action (VLA) policies leverage pretrained vision-language models (VLMs) to guide action generation for robot control. VLMs provide hierarchical visual-semantic representations that evolve across layers, from local visual geometry to abstract, language-aligned semantics; different manipulation tasks may therefore require different mixtures of layer representations. Meanwhile, the action module maintains intermediate representations that evolve throughout action computation and may provide useful information for subsequent decisions. However, existing VLA interfaces offer limited flexibility in representation access: VLM information is exposed through fixed layer assignments for each action layer, while intermediate action states are only propagated implicitly through residual streams without explicit reuse. We introduce LayerRoute, an action-conditioned representation routing interface that enables adaptive access to VLM layers and action representations. The Layer Mixture Router dynamically forms mixtures of cached VLM representations, while Action-State Reread reuses earlier action representations. Across diverse simulation and real-world benchmarks, LayerRoute consistently improves StarVLA-$π$ and $π_{0.5}$, achieving up to 7.2 gains on LIBERO Long with only 0.31% / 3.87% additional parameters. Ablation studies validate the benefit of action-conditioned layer routing, while routing analyses reveal structured allocation patterns across action layers and task settings.

Vision-language-action (VLA) models with billions of parameters now dominate the LIBERO manipulation benchmark, but the model capacity actually required by the benchmark remains unclear. We introduce MINERVA (MINimal Efficient Robotic Vision-Action policy), a family of deliberately compact visuomotor policies designed to measure this task-specific capacity floor. A 0.54M-parameter policy achieves 95.1% average success over 2,000 rollouts on the four standard LIBERO suites, only 2.4 points below the reported LeRobot $π_{0.5}$ result despite using 7,700$\times$ fewer parameters. Performance saturates near 1M parameters and collapses below 0.25M. Across broad architectural, training, and inference sweeps, only action-chunk length and vision capacity consistently exceed a $\pm$1-point training-seed band. Flow matching provides no detectable advantage over direct L1 regression across three seeds, while regression is up to 3.8$\times$ faster on GPU. A task-ID permutation probe shows that standard LIBERO instruction conditioning primarily selects among memorized tasks: changing only the task-ID mapping reduces success to near chance. The same recipe achieves 94.6% success across 89 LIBERO-90 tasks, while LIBERO-Plus perturbations reduce performance to 46--56%, with near-zero robustness to photometric shifts. The 0.54M policy replans every control step in 5--9 ms per chunk on a laptop CPU, 113$\times$ faster than SmolVLA and 1,400$\times$ faster than $π_{0.5}$, without a GPU. These results establish a first empirical estimate of LIBERO's task-specific capacity floor and motivate capacity-aware design and distillation for deployment-efficient robot policies.

Direct vision-language-action policies generate continuous robot actions efficiently, but standard behavior cloning leaves two complementary gaps: their representations are not explicitly required to describe how the scene evolves over multiple time scales, and deployment trajectories of unequal quality are often reused without separating useful dynamics from undesirable behavior. We introduce \method, a direct world-action policy that combines outcome-agnostic predictive learning with outcome-aware policy improvement. \method first retains a local fixed-offset JEPA objective and adds trajectory-relative multi-horizon transition alignment at 25%, 50%, 75%, and 100% of the remaining episode. These training-only targets require the current policy representation to preserve both local physical changes and longer-range task progress, without supplying explicit future tokens to the action head. \method then trains an independent distributional value critic on cumulative deployment trajectories, computes action-chunk-aligned $N$-step advantages, and converts them into positive, negative, or null text conditions for a flow-matching actor. Thus, every valid trajectory can teach what physically happened, while the actor is deployed only under the condition associated with relatively better actions. The multi-horizon predictor and critic are removed from online execution, preserving direct action generation from the current observation, language instruction, and proprioception. \redclaim{Across the three simulation benchmarks, \method achieves the strongest overall performance while preserving the direct actor's online execution path.}

Conventional flow-based vision-language-action (VLA) models support expressive continuous action generation but rely on multi-step refinement to produce each action chunk, increasing latency in online robot control. To address this issue, we introduce DriftingVLA, a native one-step VLA that generates a complete action chunk with a single action-expert forward pass. Rather than learning a flow field that requires iterative integration at inference, DriftingVLA uses a distribution-drifting objective to learn a direct noise-to-action-chunk mapping for one-step deployment. Since robot action dimensions carry distinct control semantics and distributional characteristics, we further introduce Per-Dimension Temporal Drifting (PDTD). PDTD treats the complete temporal trajectory of each action dimension as a separate drifting unit, enabling finer-grained modeling and shaping of dimension-specific action distributions. This per-dimension decomposition applies only to the training objective; the shared VLA model still generates the complete action chunk jointly, thereby preserving cross-dimensional dependencies. DriftingVLA achieves 98.32% success on LIBERO, 81.09% on RoboTwin 2.0, and 77.67% across six real-world single- and dual-arm tasks, outperforming the evaluated multi-step flow policy and one-step VLA baselines. Native one-step deployment also delivers a 3.36-fold speedup in action-chunk generation, eliminating iterative refinement without sacrificing control performance.

We propose VT-MUSE, a Multimodal Unified SEquential representation learning framework for visuotactilemanipulation. Existing approaches often encode visual and tactile observations independently before fusion, limiting their ability to capture fine-grained cross-modal dependencies. Moreover, most methods focus on observations at the current time step and overlook the temporal evolution of contact. VT-MUSE addresses both limitations through a two-stage representation learning framework. In Stage I, modality specific encoders are jointly adapted via cross-modal temporal alignment and masked-view consistency. In Stage II, a conditional variational latent model processes masked visual sequences together with full tactile histories. Auxiliary decoders reconstruct the masked recent visual observations and predict tactile depth changes, encouraging the latent representation to retain both global visual context and local contact dynamics. The learned representation is subsequently integrated into a lightweight Transformer policy through gated cross-attention. On the simulation benchmark, VT-MUSE outperforms the strongest baseline evaluated on all tasks by 11 percentage points and also achieves substantial improvements in real-world experiments.

Visual bottlenecks that focus policy inputs on regions of interest (ROIs) can improve data-efficient visuomotor learning by separating where to look from how to act. Many ROI interfaces rely on external spatial labels, such as gaze, object classes, or affordance annotations. Label-free alternatives often derive crops from trajectories by detecting gripper or motion events and centering a fixed crop at the projected end-effector. Such action-derived crops are useful spatial priors that require no additional labels, but they encode fixed choices about event timing, proxy points, and crop scale. When the visual evidence needed for control lies away from the end-effector or changes continuously with task progress, these crops can become misaligned. We propose Seeker, a task- and state-conditioned readout that learns attention from action. Starting from frozen DINOv3 features, Seeker iteratively updates a query with gathered visual evidence, producing progression-aware ROIs solely from action supervision. The learned ROI serves as a spatial interface for RGB cropping, mask-guided background augmentation, and point-cloud filtering. In simulation and the real world, Seeker improves data efficiency and robustness over no-crop, augmentation, and action-derived crop baselines. On real robots, Seeker raises average in-domain success from the best baseline's 48.3% to 76.7% and success under lighting/background shifts from 20.0% to 60.0%.

Dexterous manipulation promises substantially richer robot interaction with the physical world, but learning these behaviours remains constrained by the difficulty of collecting consistent, complete-task demonstrations. Unlike parallel-jaw manipulation, dexterous tasks require the operator to coordinate arm motion with precise, contact-rich finger behaviour throughout the task. We introduce NestDex, a nested policy-learning framework that reduces this burden by using learned hand skills to assist demonstration collection. The operator controls the arm and regulates the active hand skill through a single-DoF clutch, rather than directly specifying the full finger trajectory. The inner hand policy adapts its motion from the latest proprioceptive history, while a vision-language selector activates the appropriate skill for each task stage. The resulting demonstrations train a separate outer visuomotor policy that controls both the arm and hand without the inner policies at deployment. A hand-action variational autoencoder provides compact hand-action targets while retaining arm commands in joint space. Across real-world dexterous manipulation experiments, NestDex improves demonstration reliability and efficiency, and the resulting empirical evaluations support effective autonomous policy learning. Video Demo are available at project website https://aus.bot/research/nestdex.

World Action Models (WAMs) connect visual prediction with robot control, but supplying predictive context often requires expensive future-video generation. Direct policies avoid this cost but lack an explicit interface for accessing future-indexed predictive information. We introduce ForeWAM, a World Action Model that separates forecasting from rendering to expose and shape latent predictive context for efficient control. Its core mechanism, Future-KV, performs a single Video DiT prefill over the current visual latent and noise-initialized future slots, then reuses the resulting key-value states throughout action denoising. To make this context relevant to control, we introduce dynamics registers supervised by latent actions from a frozen teacher during training, encouraging representations of interaction-induced transitions. This reusable context supports a lightweight, single-layer action decoder. We evaluate ForeWAM on LIBERO, LIBERO-Plus, RoboCasa, and real-world manipulation tasks. Without additional policy-level embodied pretraining, ForeWAM improves RoboCasa success by 9.7 percentage points over Fast-WAM at the same budget of 50 demonstrations per task, reaching 59.2%. With a single-layer decoder, it achieves 77.6% success on LIBERO-Plus and reduces policy-query latency to 88.7 ms on an NVIDIA A800, delivering a 6.27-fold speedup over Fast-WAM. These results show that latent predictive computation provides useful foresight for robust, efficient control without explicit future-video generation.

Learning reliable surgical manipulation policies is bottlenecked by the scarcity of action-labeled demonstrations: teleoperated surgical robot (e.g., dVRK) trajectories with synchronized kinematics are costly to collect, while surgical tasks demand precise contact handling, long-horizon reasoning, and bimanual coordination. Endoscopic video is comparatively inexpensive and abundant relative to synchronized video--kinematics trajectories, and a natural way to exploit it is to learn world models of surgical scenes. However, existing surgical world models use video primarily for simulation or policy evaluation, and rarely translate the learned dynamics into closed-loop control. This gap raises our central question: under a fixed budget of action-labeled demonstrations, does action-free video pretraining improve closed-loop surgical manipulation? To answer it, we introduce the Surgical World-Action Model (Surgical WAM), a unified generative model built on Cosmos Policy that jointly predicts future endoscopic observations and executable surgical robot action chunks. Surgical WAM first learns surgical visual dynamics from action-free video and is then fine-tuned on the fixed action-labeled budget; at deployment, it acts as a closed-loop, receding-horizon controller that executes a short prefix of each predicted action chunk and replans from the resulting observation. On a suite of four simulated surgical manipulation tasks, video pretraining improves the average success rate from 63.5% to 77.8%, including an absolute gain of 20 percentage points on PegTransfer, with the largest improvements on contact-rich and bimanual tasks. These results demonstrate that action-free video provides transferable visual dynamics priors for learning surgical robot control with limited action supervision, positioning data-efficient video pretraining as a practical path toward scaling up surgical robot learning.