Diffusion Policy

Cooperative multi-agent reinforcement learning (MARL) involves complex agent interactions and requires effective exploration strategies. A prominent class of MARL algorithms, decentralized softmax policy gradient (DecSPG), addresses this through energy-based policy updates. In practice, however, such energy-based policies are intractable to maintain and are commonly projected onto the Gaussian policy class. In this work, we show that the limited expressiveness of Gaussian policies severely hinders exploration in DecSPG, and this limitation worsens as the number of agents grows. To address this issue, we propose decentralized diffusion policy learning (DDPL), which parameterizes each agent's policy with a denoising diffusion probabilistic model, an expressive generative model that captures multi-modal action distributions for enhanced exploration. DDPL enables efficient online training of diffusion policies via importance sampling score matching (ISSM), a novel training method with theoretical guarantee. We evaluate DDPL on representative continuous-action MARL benchmarks, including multi-agent particle environment, multi-agent MuJoCo, IsaacLab, and JAX-reimplemented StarCraft multi-agent challenge, and observe consistently improved performance.

We introduce ReflectDrive-2, a masked discrete diffusion planner with separate action expert for autonomous driving that represents plans as discrete trajectory tokens and generates them through parallel masked decoding. This discrete token space enables in-place trajectory revision: AutoEdit rewrites selected tokens using the same model, without requiring an auxiliary refinement network. To train this capability, we use a two-stage procedure. First, we construct structure-aware perturbations of expert trajectories along longitudinal progress and lateral heading directions and supervise the model to recover the original expert trajectory. We then fine-tune the full decision--draft--reflect rollout with reinforcement learning (RL), assigning terminal driving reward to the final post-edit trajectory and propagating policy-gradient credit through full-rollout transitions. Full-rollout RL proves crucial for coupling drafting and editing: under supervised training alone, inference-time AutoEdit improves PDMS by at most $0.3$, whereas RL increases its gain to $1.9$. We also co-design an efficient reflective decoding stack for the decision--draft--reflect pipeline, combining shared-prefix KV reuse, Alternating Step Decode, and fused on-device unmasking. On NAVSIM, ReflectDrive-2 achieves $91.0$ PDMS with camera-only input and $94.8$ PDMS in a best-of-6 oracle setting, while running at $31.8$ ms average latency on NVIDIA Thor.

Diffusion-based visuomotor policies perform well in robotic manipulation, yet current methods still inherit image-generation-style decoders and multi-step sampling. We revisit this design from a frequency-domain perspective. Robot action trajectories are highly smooth, with most energy concentrated in a few low-frequency discrete cosine transform modes. Under this structure, we show that the error of the optimal denoiser is bounded by the low-frequency subspace dimension and residual high-frequency energy, implying that denoising error saturates after very few reverse steps. This also suggests that action denoising requires a much simpler denoising model than image generation. Motivated by this insight, we propose Hyper-DP3 (HDP3), a pocket-scale 3D diffusion policy with a lightweight Diffusion Mixer decoder that supports two-step DDIM inference. Our synthetic experiments validate the theory and support the sufficiency of two-step denoising. Futhermore, across RoboTwin2.0, Adroit, MetaWorld, and real-world tasks, HDP3 achieves state-of-the-art performance with fewer than 1% of the parameters of prior 3D diffusion-based policies and substantially lower inference latency.

This paper introduces EnergyFlow, a framework that unifies generative action modeling with inverse reinforcement learning by parameterizing a scalar energy function whose gradient is the denoising field. We establish that under maximum-entropy optimality, the score function learned via denoising score matching recovers the gradient of the expert's soft Q-function, enabling reward extraction without adversarial training. Formally, we prove that constraining the learned field to be conservative reduces hypothesis complexity and tightens out-of-distribution generalization bounds. We further characterize the identifiability of recovered rewards and bound how score estimation errors propagate to action preferences. Empirically, EnergyFlow achieves state-of-the-art imitation performance on various manipulation tasks while providing an effective reward signal for downstream reinforcement learning that outperforms both adversarial IRL methods and likelihood-based alternatives. These results show that the structural constraints required for valid reward extraction simultaneously serve as beneficial inductive biases for policy generalization. The code is available at https://github.com/sotaagi/EnergyFlow.

Unlike chatbots, physical AI must act while the world keeps evolving. Therefore, the inter-chunk pause of synchronous executors are fatal for dynamic tasks regardless of how fast the inference is. Asynchronous execution -- thinking while acting -- is therefore a structural requirement, and real-time chunking (RTC) makes it viable by recasting chunk transitions as inpainting: freezing committed actions and consistently generating the remainder. However, RTC with flow-matching policy is structurally suboptimal: its inpainting comes from inference-time corrections rather than the base policy, yielding little pre-training benefit, specific fine-tuning, heuristic guidance, and extra computation that inflates the latency. In this work, we observe that discrete diffusion policies, which generate actions by iteratively unmasking, are natural asynchronous executors that resolve all limitations at once: they are fine-tuning free since inpainting is their native operation, while early stopping further provides adaptive guidance and reduces inference cost. We propose DiscreteRTC, which replaces external corrections with native unmasking, and show on dynamic simulated benchmarks and real-world dynamic manipulation tasks that it achieves higher success rates than continuous RTC and other baselines. In summary, DiscreteRTC is simpler to implement with 0 lines of additional code to enable async inpainting, faster at inference with only ~0.7 computation compared with generating actions from scratch, and better at execution with 65% higher success rate in real-world hockey defend task compared with flow-matching RTC, and 30% higher compared with training-time flow-matching RTC. More visualizations are on https://outsider86.github.io/DiscreteRTCSite/.

Contact-rich manipulation is central to many everyday human activities, requiring continuous adaptation to contact uncertainty and external disturbances through multi-modal perception, particularly vision and tactile feedback. While imitation learning has shown strong potential for learning complex manipulation behaviors, most existing approaches rely on action chunking, which fundamentally limits their ability to react to unforeseen observations during execution. This limitation becomes especially critical in contact-rich scenarios, where physical uncertainty and high-frequency tactile feedback demand rapid, reactive control. To address this challenge, we propose Tube Diffusion Policy (TDP), a novel reactive visual-tactile policy learning framework that bridges diffusion-based imitation learning with tube-based feedback control. By leveraging the expressive power of generative models, TDP learns an observation-conditioned feedback flow around nominal action chunks, forming an action tube that enables fast and adaptive reactions during execution. We evaluate TDP on the widely used Push-T benchmark and three additional challenging visual-tactile dexterous manipulation tasks. Across all benchmarks, TDP consistently outperforms state-of-the-art imitation learning baselines. Two real-world experiments further validate its robust reactivity under contact uncertainty and external disturbances. Moreover, the step-wise correction mechanism enabled by action tube significantly reduces the required denoising steps, making TDP well suited for real-time, high-frequency feedback control in contact-rich manipulation.

Bridging high-level semantic understanding with low-level physical control remains a persistent challenge in embodied intelligence, stemming from the fundamental spatiotemporal scale mismatch between cognition and action. Existing generative VLA policies typically adopt a "Generation-from-Noise" paradigm, which disregards this disparity, leading to representation inefficiency and weak condition alignment during optimization. In this work, we propose ResVLA, an architecture that shifts the paradigm to "Refinement-from-Intent." Recognizing that robotic motion naturally decomposes into global intent and local dynamics, ResVLA utilizes spectral analysis to decouple control into a deterministic low-frequency anchor and a stochastic high-frequency residual. By anchoring the generative process on the predicted intent, our model focuses strictly on refining local dynamics via a residual diffusion bridge. Extensive simulation experiments show that ResVLA achieves competitive performance, strong robustness to language and robot embodiment perturbations, and faster convergence than standard generative baselines. ResVLA also demonstrates strong performance in real-world robot experiments.

Some of the most performant reinforcement learning algorithms today can be prohibitively expensive as they use test-time scaling methods such as sampling multiple action candidates and selecting the best one. In this work, we propose FASTER, a method for getting the benefits of sampling-based test-time scaling of diffusion-based policies without the computational cost by tracing the performance gain of action samples back to earlier in the denoising process. Our key insight is that we can model the denoising of multiple action candidates and selecting the best one as a Markov Decision Process (MDP) where the goal is to progressively filter action candidates before denoising is complete. With this MDP, we can learn a policy and value function in the denoising space that predicts the downstream value of action candidates in the denoising process and filters them while maximizing returns. The result is a method that is lightweight and can be plugged into existing generative RL algorithms. Across challenging long-horizon manipulation tasks in online and batch-online RL, FASTER consistently improves the underlying policies and achieves the best overall performance among the compared methods. Applied to a pretrained VLA, FASTER achieves the same performance while substantially reducing training and inference compute requirements. Code is available at https://github.com/alexanderswerdlow/faster .

Trajectory Optimization (TO) solvers exploit known system dynamics to compute locally optimal trajectories through iterative improvements. A downside is that each new problem instance is solved independently; therefore, convergence speed and quality of the solution found depend on the initial trajectory proposed. To improve efficiency, a natural approach is to warm-start TO with initial guesses produced by a learned policy trained on trajectories previously generated by the solver. Diffusion-based policies have recently emerged as expressive imitation learning models, making them promising candidates for this role. Yet, a counterintuitive challenge comes from the local optimality of TO demonstrations: when a policy is rolled out, small non-optimal deviations may push it into situations not represented in the training data, triggering compounding errors over long horizons. In this work, we focus on learning-based warm-starting for gradient-based TO solvers that also provide feedback gains. Exploiting this specificity, we derive a first-order loss for Sobolev learning of diffusion-based policies using both trajectories and feedback gains. Through comprehensive experiments, we demonstrate that the resulting policy avoids compounding errors, and so can learn from very few trajectories to provide initial guesses reducing solving time by $2\times$ to $20 \times$. Incorporating first-order information enables predictions with fewer diffusion steps, reducing inference latency.

Diffusion policies are becoming mainstream in robotic manipulation but suffer from hard negative class imbalance due to uniform sampling and lack of sample difficulty awareness, leading to slow training convergence and frequent inference timeout failures. We propose VADF (Vision-Adaptive Diffusion Policy Framework), a vision-driven dual-adaptive framework that significantly reduces convergence steps and achieves early success in inference, with model-agnostic design enabling seamless integration into any diffusion policy architecture. During training, we introduce Adaptive Loss Network (ALN), a lightweight MLP-based loss predictor that quantifies per-step sample difficulty in real time. Guided by hard negative mining, it performs weighted sampling to prioritize high-loss regions, enabling adaptive weight updates and faster convergence. In inference, we design the Hierarchical Vision Task Segmenter (HVTS), which decomposes high-level task instructions into multi-stage low-level sub-instructions based on visual input. It adaptively segments action sequences into simple and complex subtasks by assigning shorter noise schedules with longer direct execution sequences to simple actions, and longer noise steps with shorter execution sequences to complex ones, thereby dramatically reducing computational overhead and significantly improving the early success rate.

3D policy learning promises superior generalization and cross-embodiment transfer, but progress has been hindered by training instabilities and severe overfitting, precluding the adoption of powerful 3D perception models. In this work, we systematically diagnose these failures, identifying the omission of 3D data augmentation and the adverse effects of Batch Normalization as primary causes. We propose a new architecture coupling a scalable transformer-based 3D encoder with a diffusion decoder, engineered specifically for stability at scale and designed to leverage large-scale pre-training. Our approach significantly outperforms state-of-the-art 3D baselines on challenging manipulation benchmarks, establishing a new and robust foundation for scalable 3D imitation learning. Project Page: https://r3d-policy.github.io/

While end-to-end Vision-Language-Action (VLA) models offer a promising paradigm for robotic manipulation, fine-tuning them on narrow control data often compromises the profound reasoning capabilities inherited from their base Vision-Language Models (VLMs). To resolve this fundamental trade-off, we propose HiVLA, a visual-grounded-centric hierarchical framework that explicitly decouples high-level semantic planning from low-level motor control. In high-level part, a VLM planner first performs task decomposition and visual grounding to generate structured plans, comprising a subtask instruction and a precise target bounding box. Then, to translate this plan into physical actions, we introduce a flow-matching Diffusion Transformer (DiT) action expert in low-level part equipped with a novel cascaded cross-attention mechanism. This design sequentially fuses global context, high-resolution object-centric crops and skill semantics, enabling the DiT to focus purely on robust execution. Our decoupled architecture preserves the VLM's zero-shot reasoning while allowing independent improvement of both components. Extensive experiments in simulation and the real world demonstrate that HiVLA significantly outperforms state-of-the-art end-to-end baselines, particularly excelling in long-horizon skill composition and the fine-grained manipulation of small objects in cluttered scenes.

Achieving precise, versatile whole-body character control in physics-based animation remains challenging. Recent diffusion-based policies generate rich and expressive motions but typically rely on gradient-based test-time guidance to satisfy task objectives, which is slow and can reduce robustness. We introduce NaP-Control (Navigating Diffusion Prior for Versatile and Fast Character Control), abbreviated as NaP. Our method uses reinforcement learning to manipulate the latent noise of a task-agnostic diffusion policy prior, steering it toward task-specific behaviors for fast, robust control with high motion fidelity. In contrast to methods that rely solely on offline training, NaP interacts with the environment during training to correct motions and optimize task rewards, improving success rates and enabling adaptation to challenging scenarios. By directly predicting task-optimized diffusion noise, NaP eliminates iterative guidance during denoising and enables efficient inference. Experiments show that NaP attains higher success rates and faster inference while preserving natural motion across diverse tasks.

This paper addresses a fundamental problem of visuomotor policy learning for robotic manipulation: how to enhance robustness in out-of-distribution execution errors or dynamically re-routing trajectories, where the model relies solely on the original expert demonstrations for training. We introduce the Referring-Aware Visuomotor Policy (ReV), a closed-loop framework that can adapt to unforeseen circumstances by instantly incorporating sparse referring points provided by a human or a high-level reasoning planner. Specifically, ReV leverages the coupled diffusion heads to preserve standard task execution patterns while seamlessly integrating sparse referring via a trajectory-steering strategy. Upon receiving a specific referring point, the global diffusion head firstly generates a sequence of globally consistent yet temporally sparse action anchors, while identifies the precise temporal position for the referring point within this sequence. Subsequently, the local diffusion head adaptively interpolates adjacent anchors based on the current temporal position for specific tasks. This closed-loop process repeats at every execution step, enabling real-time trajectory replanning in response to dynamic changes in the scene. In practice, rather than relying on elaborate annotations, ReV is trained only by applying targeted perturbations to expert demonstrations. Without any additional data or fine-tuning scheme, ReV achieve higher success rates across challenging simulated and real-world tasks.

We consider constrained ergodic resource optimization in wireless networks with graph-structured interference. We train a diffusion model policy to match expert conditional distributions over resource allocations. By leveraging a primal-dual (expert) algorithm, we generate primal iterates that serve as draws from the corresponding expert conditionals for each training network instance. We view the allocations as stochastic graph signals supported on known channel state graphs. We implement the diffusion model architecture as a U-Net hierarchy of graph neural network (GNN) blocks, conditioned on the channel states and additional node states. At inference, the learned generative model amortizes the iterative expert policy by directly sampling allocation vectors from the near-optimal conditional distributions. In a power-control case study, we show that time-sharing the generated power allocations achieves near-optimal ergodic sum-rate utility and near-feasible ergodic minimum-rates, with strong generalization and transferability across network states.

Diffusion policies effectively model multimodal action distributions for robotic manipulation, but their iterative denoising requires tens to hundreds of network function evaluations (NFEs) for each control prediction, limiting their applicability to high-frequency closed-loop control and online reinforcement learning (RL). We present a two-stage framework for native one-step generative policies that transfers iterative refinement from inference to training. First, Drift-Based Policy (DBP) uses a fixed-point drifting objective to internalize corrective dynamics into the model parameters, producing multimodal action chunks with a single network evaluation by design. Second, Drift-Based Policy Optimization (DBPO) augments the pretrained backbone with a stochastic interface that provides exact conditional rollout likelihoods for PPO-style on-policy updates while preserving 1-NFE deployment. On the 12-task Diffusion Policy suite, DBP improves the average success rate from 0.79 to 0.83 while reducing inference from 100 NFEs to 1. Across 37 point-cloud manipulation tasks, DBP achieves an average success rate of 88.4%, surpassing the leading 1-NFE baseline OMP at 82.3%. DBPO further improves pretrained one-step policies through stable online fine-tuning on RoboMimic and D4RL. On a physical dual-arm UR5 platform, DBP achieves 123/150 successes (82%) with an average end-to-end latency of 9.5 ms, compared with MP1's 89/150 successes (59%) under the same setup. Code is available at https://github.com/YuxuanGao0822/DBPO.

Active multi-target tracking requires a mobile robot to balance exploration for undetected targets with exploitation of uncertain tracked ones. Diffusion policies have emerged as a powerful approach for capturing diverse behavioral strategies by learning action sequences from expert demonstrations. However, existing methods implicitly select among strategies through the denoising process, without uncertainty quantification over which strategy to execute. We formulate expert selection for diffusion policies as an offline contextual bandit problem and propose a Bayesian framework for pessimistic, uncertainty-aware strategy selection. A multi-head Variational Bayesian Last Layer (VBLL) model predicts the expected tracking performance of each expert strategy given the current belief state, providing both a point estimate and predictive uncertainty. Following the pessimism principle for offline decision-making, a Lower Confidence Bound (LCB) criterion then selects the expert whose worst-case predicted performance is best, avoiding overcommitment to experts with unreliable predictions. The selected expert conditions a diffusion policy to generate corresponding action sequences. Experiments on simulated indoor tracking scenarios demonstrate that our approach outperforms both the base diffusion policy and standard gating methods, including Mixture-of-Experts selection and deterministic regression baselines.

A central challenge in mobile manipulation is preserving multiple plausible action modes while remaining reactive during execution. For example, grasping a bottle in a cluttered scene can be achieved through multiple physically feasible base-arm motion plans. Deterministic VLA policies trained with $L_1$ regression are efficient, but they tend to average across valid modes and may produce physically invalid actions. Diffusion-based generative policies can better represent multimodal action distributions, yet their proposals may be weakly constrained when generated from pure noise, and iterative sampling can increase inference cost. Action chunking helps amortize inference, but it also introduces partially open-loop execution, allowing small mismatches to accumulate into drift. We present MobiGuide, a generative VLA model for efficient mobile manipulation, where action generation is guided by a structured trajectory vocabulary. Constructed from expert demonstrations, this vocabulary encodes plausible whole-body motion patterns and provides structured priors for feasible action generation. Given the current observation and instruction, MobiGuide locally refines these reference trajectories into context-compatible action proposals and selects the most suitable trajectory for execution. This preserves multimodal action candidates while keeping generation constrained around feasible expert-like motion modes. To mitigate chunk-induced drift, we further introduce a lightweight residual correction module that performs high-frequency, per-step adjustments during rollout. Across diverse mobile manipulation tasks, MobiGuide improves success and stability under execution disturbances while maintaining low-latency inference.

Vision-based policies have achieved a good performance in robotic manipulation due to the accessibility and richness of visual observations. However, purely visual sensing becomes insufficient in contact-rich and force-sensitive tasks where force/torque (F/T) signals provide critical information about contact dynamics, alignment, and interaction quality. Although various strategies have been proposed to integrate vision and F/T signals, including auxiliary prediction objectives, mixture-of-experts architectures, and contact-aware gating mechanisms, a comparison of these approaches remains lacking. In this work, we provide a controlled comparison of different F/T-vision integration strategies within diffusion-based manipulation policies. In addition, we propose an adaptive integration strategy that ignores F/T signals during non-contact phases while adaptively leveraging both vision and torque information during contact. Experimental results demonstrate that our method outperforms the strongest baseline by 14% in success rate, highlighting the importance of contact-aware multimodal fusion for robotic manipulation.

Multi-agent reinforcement learning (MARL) provides a promising solution for cooperative target tracking in networks of autonomous underwater vehicles (AUVs). However, existing methods still face three major challenges: 1) policy non-stationarity caused by concurrent updates among multiple agents; 2) inefficient policy learning caused by the heterogeneous quality of experiences accumulated during exploration; and 3) policy drift between stochastic exploration and deterministic execution under dynamic underwater disturbances. To address these challenges, this paper develops a four-layer hierarchical MARL architecture comprising global training scheduling, multi-agent coordination, local policy generation, and real-time action execution. Building on this architecture, we propose a Supervised Diffusion-Aided MARL (SDA-MARL) algorithm with three closely coupled mechanisms. First, a dual-decision policy integrates a diffusion-based generative branch with a Deep Deterministic Policy Gradient (DDPG) branch, while segregated experience pools reduce training interference between the two branches. Second, a supervised sample-selection mechanism identifies high-quality tracking transitions and uses their actions to guide reverse diffusion, enabling the generative policy to concentrate on effective regions of the action space. Third, a behavioral-cloning loss transfers diffusion-generated actions to the deterministic DDPG Actor, thereby aligning exploration with execution and suppressing policy drift. Experiments conducted in six-degree-of-freedom underwater environments across multiple AUV-target configurations show that SDA-MARL achieves faster convergence, higher tracking accuracy, more consistent inter-AUV velocities, and shorter tracking paths than the compared MARL methods.

Most existing vision-language-action (VLA) models for robotic manipulation lack progress awareness, typically relying on hand-crafted heuristics for task termination. This limitation is particularly severe in long-horizon tasks involving cascaded sub-goals. In this work, we investigate the estimation and integration of task progress, proposing a novel model named {\textbf \vla}. Our technical contributions are twofold: (1) robust progress estimation: We pre-train a progress estimator on large-scale, unsupervised video-text robotic datasets. This estimator achieves a low prediction residual (0.07 on a scale of $[0, 1]$) in simulation and demonstrates zero-shot generalization to unseen real-world samples, and (2) differentiable progress guidance: We introduce an inverse dynamics world model that maps predicted action tokens into future latent visual states. These latents are then processed by the progress estimator; by applying a maximal progress regularization, we establish a differentiable pipeline that provides progress-piloted guidance to refine action tokens. Extensive experiments on the CALVIN and LIBERO benchmarks, alongside real-world robot deployment, consistently demonstrate substantial improvements in success rates and generalization over strong baselines.

Thanks to their remarkable flexibility, diffusion models and flow models have emerged as promising candidates for policy representation. However, efficient reinforcement learning (RL) upon these policies remains a challenge due to the lack of explicit log-probabilities for vanilla policy gradient estimators. While numerous attempts have been proposed to address this, the field lacks a unified perspective to reconcile these seemingly disparate methods, thus hampering ongoing development. In this paper, we bridge this gap by introducing a comprehensive taxonomy for RL algorithms with diffusion/flow policies. To support reproducibility and agile prototyping, we introduce a modular, JAX-based open-source codebase that leverages JIT-compilation for high-throughput training. Finally, we provide systematic and standardized benchmarks across Gym-Locomotion, DeepMind Control Suite, and IsaacLab, offering a rigorous side-by-side comparison of diffusion-based methods and guidance for practitioners to choose proper algorithms based on the application. Our work establishes a clear foundation for understanding and algorithm design, a high-efficiency toolkit for future research in the field, and an algorithmic guideline for practitioners in generative models and robotics. Our code is available at https://github.com/typoverflow/flow-rl.

Shared autonomy combines human user and AI copilot actions to control complex systems such as robotic arms. When a task is challenging, requires high dimensional control, or is subject to corruption, shared autonomy can significantly increase task performance by using a trained copilot to effectively correct user actions in a manner consistent with the user's goals. To significantly improve the performance of shared autonomy, we introduce Diffusion Sequence Copilots (DiSCo): a method of shared autonomy with diffusion policy that plans action sequences consistent with past user actions. DiSCo seeds and inpaints the diffusion process with user-provided actions with hyperparameters to balance conformity to expert actions, alignment with user intent, and perceived responsiveness. We demonstrate that DiSCo substantially improves task performance in simulated driving and robotic arm tasks. Project website: https://sites.google.com/view/disco-shared-autonomy/

Diffusion policies excel at visuomotor control but often fail catastrophically under severe out-of-distribution (OOD) disturbances, such as unexpected object displacements or visual corruptions. To address this vulnerability, we introduce the Dream Diffusion Policy (DDP), a framework that deeply integrates a diffusion world model into the policy's training objective via a shared 3D visual encoder. This co-optimization endows the policy with robust state-prediction capabilities. When encountering sudden OOD anomalies during inference, DDP detects the real-imagination discrepancy and actively abandons the corrupted visual stream. Instead, it relies on its internal "imagination" (autoregressively forecasted latent dynamics) to safely bypass the disruption, generating imagined trajectories before smoothly realigning with physical reality. Extensive evaluations demonstrate DDP's exceptional resilience. Notably, DDP achieves a 73.8% OOD success rate on MetaWorld (vs. 23.9% without predictive imagination) and an 83.3% success rate under severe real-world spatial shifts (vs. 3.3% without predictive imagination). Furthermore, as a stress test, DDP maintains a 76.7% real-world success rate even when relying entirely on open-loop imagination post-initialization.

Striking a balance between efficiency and transparent motion is a core challenge in human-robot collaboration, as highly expressive movements often incur unnecessary time and energy costs. In collaborative environments, legibility allows a human observer a better understanding of the robot's actions, increasing safety and trust. However, these behaviors result in sub-optimal and exaggerated trajectories that are redundant in low-ambiguity scenarios where the robot's goal is already obvious. To address this trade-off, we propose Style-Conditioned Diffusion Policy (SCDP), a modular framework that constrains the trajectory generation of a pre-trained diffusion model toward either legibility or efficiency based on the environment's configuration. Our method utilizes a post-training pipeline that freezes the base policy and trains a lightweight scene encoder and conditioning predictor to modulate the diffusion process. At inference time, an ambiguity detection module activates the appropriate conditioning, prioritizing expressive motion only for ambiguous goals and reverting to efficient paths otherwise. We evaluate SCDP on manipulation and navigation tasks, and results show that it enhances legibility in ambiguous settings while preserving optimal efficiency when legibility is unnecessary, all without retraining the base policy.

Generative robot policies trained on demonstrations using behavior cloning often learn actions that are sub-optimal or misaligned with respect to the downstream task. Policy improvement approaches aim to bridge this gap and improve the cumulative reward with minimal interventions on the pre-trained policy. However, an impediment to their practical deployment is the requirement of cumbersome hyperparameter tuning specific to model or task families. In this work, we present an episodic, derivative-free latent policy improvement approach that works with little to no changes across model scales from MLPs to VLAs. We demonstrate that the performance of a pretrained, frozen diffusion or flow matching policy can be improved with respect to a downstream reward by swapping the sampling of initial noise from the prior distribution (typically isotropic Gaussian) with a well-chosen, constant initial noise input---a golden ticket. We show the prevalence of golden tickets by improving the policy performance of $46$ out of $51$ tasks across manipulation benchmarks, with absolute improvements in success rate by up to $79\%$ for simulated tasks, and $28\%$ within $60$ search episodes for real-world tasks. Further, we find that the versatility of our approach opens up new avenues such as simultaneous policy improvement for multiple downstream rewards. Project webpage: https://lottery-tickets.rai-inst.com/.

Robotic contact-rich and fine-grained manipulation remains a significant challenge due to complex interaction dynamics and the competing requirements of multi-timescale control. While current visual imitation learning methods excel at long-horizon planning, they often fail to perceive critical interaction cues like friction variations or incipient slip, and struggle to balance global task coherence with local reactive feedback. To address these challenges, we propose M2-ResiPolicy, a novel Master-Micro residual control architecture that synergizes high-level action guidance with low-level correction. The framework consists of a Master-Guidance Policy (MGP) operating at 10 Hz, which generates temporally consistent action chunks via a diffusion-based backbone and employs a tactile-intensity-driven adaptive fusion mechanism to dynamically modulate perceptual weights between vision and touch. Simultaneously, a high-frequency (60 Hz) Micro-Residual Corrector (MRC) utilizes a lightweight GRU to provide real-time action compensation based on TCP wrench feedback. This policy is further integrated with a force-mixed PBIC execution layer, effectively regulating contact forces to ensure interaction safety. Experiments across several demanding tasks including fragile object grasping and precision insertion, demonstrate that M2-ResiPolicy significantly outperforms standard Diffusion Policy (DP) and state-of-the-art Reactive Diffusion Policy (RDP), achieving a 93% damage-free success rate in chip grasping and superior force regulation stability.

Generalist robot policies built upon 2D visual representations excel at semantic reasoning but inherently lack the explicit 3D spatial awareness required for high-precision tasks. Existing 3D integration methods struggle to bridge this gap due to the structural irregularity of sparse point clouds and the geometric distortion introduced by multi-view orthographic rendering. To overcome these barriers, we present ReMAP-DP, a novel framework synergizing standardized perspective reprojection with a structure-aware dual-stream diffusion policy. By coupling the re-projected views with pixel-aligned PointMaps, our dual-stream architecture leverages learnable modality embeddings to fuse frozen semantic features and explicit geometric descriptors, ensuring precise implicit patch-level alignment. Extensive experiments across simulation and real-world environments demonstrate ReMAP-DP's superior performance in diverse manipulation tasks. On RoboTwin 2.0, it attains a 59.3% average success rate, outperforming the DP3 baseline by +6.6%. On ManiSkill 3, our method yields a 28% improvement over DP3 on the geometrically challenging Stack Cube task. Furthermore, ReMAP-DP exhibits remarkable real-world robustness, executing high-precision and dynamic manipulations with superior data efficiency from only a handful of demonstrations. Project page is available at: https://icr-lab.github.io/ReMAP-DP/

Humanoid loco-manipulation requires coordinated task-space motion planning with stable loco-manipulation command tracking under complex robot-environment dynamics and long-horizon tasks. While diffusion policies (DPs) show promise for learning from demonstrations, deploying them on humanoids poses critical challenges: the motion planner trained offline is decoupled from the loco-manipulation controller, leading to poor command tracking, compounding distribution shift, and task failures. The common approach of scaling demonstration data is prohibitively expensive for high-dimensional humanoid systems. To address this challenge, we present REFINE-DP (REinforcement learning FINE-tuning of Diffusion Policy), a hierarchical framework that jointly optimizes a DP motion planner and an RL-based loco-manipulation controller. The DP is fine-tuned via a PPO-based diffusion policy gradient to improve task success rate, while the controller is simultaneously updated to accurately track the planner's evolving command distribution, reducing the distributional mismatch that degrades motion quality. We validate REFINE-DP on a humanoid robot performing loco-manipulation tasks, including door traversal and long-horizon object transport. REFINE-DP achieves an over 90% success rate in simulation, even in out-of-distribution cases not seen in the pre-training data, and enables real-world execution without privileged state information. Our proposed method substantially outperforms pre-trained DP baselines and demonstrates that RL fine-tuning is key to reliable humanoid loco-manipulation. https://refine-dp.github.io/REFINE-DP/

Diffusion-based robot navigation policies trained on large-scale imitation learning datasets, can generate multi-modal trajectories directly from the robot's visual observations, bypassing the traditional localization-mapping-planning pipeline and achieving strong zero-shot generalization. However, their performance remains constrained by the coverage of offline datasets, and when deployed in unseen settings, distribution shift often leads to accumulated trajectory errors and safety-critical failures. Adapting diffusion policies with reinforcement learning is challenging because their iterative denoising structure hinders effective gradient backpropagation, while also making the training of an additional value network computationally expensive and less stable. To address these issues, we propose a reinforcement learning fine-tuning framework tailored for diffusion-based navigation. The method leverages the inherent multi-trajectory sampling mechanism of diffusion models and adopts Group Relative Policy Optimization (GRPO), which estimates relative advantages across sampled trajectories without requiring a separate value network. To preserve pretrained representations while enabling adaptation, we freeze the visual encoder and selectively update the higher decoder layers and action head, enhancing safety-aware behaviors through online environmental feedback. On the PointGoal task in Isaac Sim, our approach improves the Success Rate from 52.0% to 58.7% and SPL from 0.49 to 0.54 on unseen scenes, while reducing collision frequency. Additional experiments show that the fine-tuned policy transfers zero-shot to a real quadruped platform and maintains stable performance in geometrically out-of-distribution environments, suggesting improved adaptability and safe generalization to new domains.