Diffusion Policy

Action chunking is widely used in generative visuomotor policies, yet the recurring execution discontinuities at chunk boundaries still lack a mechanistic explanation. This paper treats chunk-boundary artifact as an analyzable mechanism variable. We first show that successful and failed episodes separate stably on artifact metrics. We then show that, in stochastic action-chunked policies, fixing the observation context and changing only latent noise is sufficient to modulate artifact systematically. On the same Diffusion Policy checkpoint, comparisons among DDPM, zero-variance DDPM, and DDIM further show that this local controllability depends on whether the information path from initial noise to action output remains intact. Finally, from controlled interventions at fixed local execution states, we find that artifact changes can carry through to final outcome, and that the preferred direction can reverse even within the same task: some contexts achieve higher success under lower artifact, whereas others achieve higher success under higher artifact. In a representative high-artifact-favoring key context selected by held-out matched-continuation validation, success rate increases from 0.033 to 0.717. These results show that chunk-boundary artifact is not a mere execution-side by-product, but a variable in noise space that can be attributed, controlled, and mechanistically linked to task outcome.

Diffusion policies have shown to be very efficient at learning complex, multi-modal behaviors for robotic manipulation. However, errors in generated action sequences can compound over time which can potentially lead to failure. Some approaches mitigate this by augmenting datasets with expert demonstrations or learning predictive world models which might be computationally expensive. We introduce Performance Predictive Guidance (PPGuide), a lightweight, classifier-based framework that steers a pre-trained diffusion policy away from failure modes at inference time. PPGuide makes use of a novel self-supervised process: it uses attention-based multiple instance learning to automatically estimate which observation-action chunks from the policy's rollouts are relevant to success or failure. We then train a performance predictor on this self-labeled data. During inference, this predictor provides a real-time gradient to guide the policy toward more robust actions. We validated our proposed PPGuide across a diverse set of tasks from the Robomimic and MimicGen benchmarks, demonstrating consistent improvements in performance.

Vision-Language-Action (VLA) models have emerged as a promising paradigm for robot learning, but their representations are still largely inherited from static image-text pretraining, leaving physical dynamics to be learned from comparatively limited action data. Generative video models, by contrast, encode rich spatiotemporal structure and implicit physics, making them a compelling foundation for robotic manipulation. But their potentials are not fully explored in the literature. To bridge the gap, we introduce DiT4DiT, an end-to-end Video-Action Model that couples a video Diffusion Transformer with an action Diffusion Transformer in a unified cascaded framework. Instead of relying on reconstructed future frames, DiT4DiT extracts intermediate denoising features from the video generation process and uses them as temporally grounded conditions for action prediction. We further propose a dual flow-matching objective with decoupled timesteps and noise scales for video prediction, hidden-state extraction, and action inference, enabling coherent joint training of both modules. Across simulation and real-world benchmarks, DiT4DiT achieves state-of-the-art results, reaching average success rates of 98.6% on LIBERO and 50.8% on RoboCasa GR1 while using substantially less training data. On the Unitree G1 robot, it also delivers superior real-world performance and strong zero-shot generalization. Importantly, DiT4DiT improves sample efficiency by over 10x and speeds up convergence by up to 7x, demonstrating that video generation can serve as an effective scaling proxy for robot policy learning. We release code and models at https://dit4dit.github.io/.

In recent years, Behavior Cloning (BC) has become one of the most prevalent methods for learning manipulation from human demonstrations. Despite their successes, BC policies are often brittle and struggle with precise manipulation. To overcome these issues, we propose UF-OPS, an Update-Free On-Policy Steering method that enables the robot to predict the success likelihood of its actions and adapt its strategy at execution time. We accomplish this by training verifier functions using policy rollout data obtained during an initial evaluation of the policy. These verifiers are subsequently used to steer the base policy toward actions with a higher likelihood of success. Our method improves the performance of black-box diffusion policies, without changing the base parameters, making it lightweight and flexible. We present results from both simulation and real-world data and achieve an average 49% improvement in success rate over the base policy across 5 real tasks.

A commonly used family of RL algorithms for diffusion policies conducts softmax reweighting over samples from the behavior policy, which often induces an overgreedy policy and fails to utilize feedback from negative samples. In this work, we introduce GeMPO, a simple and unified framework that generalizes reweighting scheme in diffusion RL from softmax to general monotonic functions. GeMPO revisits diffusion RL via a measure matching perspective: First, we construct a virtual target policy measure via solving a regularized policy optimization objective; Second, we minimize the divergence between the current policy and this target measure through reweighted flow matching. This formulation offers two key advantages: i) It extends weight design beyond traditional exponential reweighting, allowing it to be tailored to diverse reward landscapes; and ii) by relaxing the non-negativity constraint on the target measure, our framework provides a principled justification for negative reweighting. We provide interpretations of how negative reweighting actively repels the policy from suboptimal actions and thus facilitates exploration. Extensive empirical evaluations demonstrate that GeMPO achieves competitive or superior performance by leveraging these flexible weighting schemes, and we provide practical guidelines for selecting reweighting methods in practice.

Distilling humanoid locomotion control from offline datasets into deployable policies remains a challenge, as existing methods rely on privileged full-body states that require complex and often unreliable state estimation. We present Sensor-Conditioned Diffusion Policies (SCDP) that enables humanoid locomotion using only onboard sensors, eliminating the need for explicit state estimation. SCDP decouples sensing from supervision through mixed-observation training: diffusion model conditions on sensor histories while being supervised to predict privileged future state-action trajectories, enforcing the model to infer the motion dynamics under partial observability. We further develop restricted denoising, context distribution alignment, and context-aware attention masking to encourage implicit state estimation within the model and to prevent train-deploy mismatch. We validate SCDP on velocity-commanded locomotion and motion reference tracking tasks. In simulation, SCDP achieves near-perfect success on velocity control (99-100%) and 93% tracking success in AMASS test set, performing comparable to privileged baselines while using only onboard sensors. Finally, we deploy the trained policy on a real G1 humanoid at 50 Hz, demonstrating robust real robot locomotion without external sensing or state estimation.

Contact-rich dexterous manipulation with multi-finger hands remains an open challenge in robotics because task success depends on multi-point contacts that continuously evolve and are highly sensitive to object geometry, frictional transitions, and slip. Recently, tactile-informed manipulation policies have shown promise. However, most use tactile signals as additional observations rather than modeling contact state or how their action outputs interact with low-level controller dynamics. We present Contact-Grounded Policy (CGP), a visuotactile policy that grounds multi-point contacts by predicting coupled trajectories of actual robot state and tactile feedback, and using a learned contact-consistency mapping to convert these predictions into executable target robot states for a compliance controller. CGP consists of two components: (i) a conditional diffusion model that forecasts future robot state and tactile feedback in a compressed latent space, and (ii) a learned contact-consistency mapping that converts the predicted robot state-tactile pair into executable targets for a compliance controller, enabling it to realize the intended contacts. We evaluate CGP using a physical four-finger Allegro V5 hand with Digit360 fingertip tactile sensors, and a simulated five-finger Tesollo DG-5F hand with dense whole-hand tactile arrays. Across a range of dexterous tasks including in-hand manipulation, delicate grasping, and tool use, CGP outperforms visuomotor and visuotactile diffusion-policy baselines.

Hierarchical policies decompose language-conditioned long-horizon robotic manipulation into a high-level planner and a low-level controller. However, effective coordination between HL and LL requires that both components operate on compatible subgoal distributions. We propose ORCHID, a self-training framework that enables stable online improvement of hierarchical diffusion policies by aligning planning and control through iterative refinement. By filtering policy samples via environment feedback, ORCHID identifies trajectories where the planner and controller are jointly successful and distills them back into both modules via supervised learning. This process induces a bidirectional co-adaptation: the planner grounds its subgoals in the actual reaching capabilities of the controller, while the controller specializes in the trajectory structures the planner produces. By relying on supervised distillation of filtered on-policy samples, ORCHID avoids the instability typical of online hierarchical gradient-based RL training with diffusion models. On the CALVIN benchmark, ORCHID allows a lightweight, initially weak model to outperform pure offline methods, including a Vision-Language-Action model twice its size.

Imitation Learning (IL) enables robots to acquire manipulation skills from expert demonstrations. Diffusion Policy (DP) models multi-modal expert behaviors but degrades when naively increasing stacked observation horizons, limiting long-horizon manipulation. We propose Self-Evolving Gated Attention (SEGA), a temporal module that maintains a time-evolving latent state via gated attention, enabling efficient recurrent updates that accumulate long-term context into a compact latent representation while filtering irrelevant temporal information. Integrating SEGA into DP yields Self-Evolving Diffusion Policy (SeedPolicy), which resolves the temporal modeling bottleneck and extends the effective temporal horizon with moderate overhead. On the RoboTwin 2.0 benchmark with 50 manipulation tasks, SeedPolicy outperforms DP and other IL baselines. Averaged across both CNN and Transformer backbones, SeedPolicy achieves 36.8% relative improvement in clean settings and 169% relative improvement in randomized challenging settings over the DP. Compared to vision-language-action models such as RDT with 1.2B parameters, SeedPolicy achieves stronger performance in the clean setting with one to two orders of magnitude fewer parameters, demonstrating strong efficiency. These results establish SeedPolicy as a state-of-the-art imitation learning method for long-horizon robotic manipulation. Code is available at: https://github.com/Youqiang-Gui/SeedPolicy.

Imitation learning from human demonstrations has achieved significant success in robotic control, yet most visuomotor policies still condition on single-step observations or short-context histories, making them struggle with non-Markovian tasks that require long-term memory. Simply enlarging the context window incurs substantial computational and memory costs and encourages overfitting to spurious correlations, leading to catastrophic failures under distribution shift and violating real-time constraints in robotic systems. By contrast, humans can compress important past experiences into long-term memories and exploit them to solve tasks throughout their lifetime. In this paper, we propose VPWEM, a non-Markovian visuomotor policy equipped with working and episodic memories. VPWEM retains a sliding window of recent observation embeddings as short-term working memory, and introduces a Transformer-based contextual memory compressor that recursively converts out-of-window observations into a fixed number of episodic memory embeddings. The compressor uses self-attention over a cache of past summary embeddings and cross-attention over a cache of historical observations, and is trained jointly with the policy. We instantiate VPWEM on diffusion policies to exploit both short-term and episode-wide information for action generation with nearly constant memory and computation per step. Experiments demonstrate that VPWEM outperforms state-of-the-art baselines including diffusion policies and vision-language-action (VLA) models by more than 20% on the memory-intensive manipulation tasks in MIKASA and achieves an average 5% improvement on the mobile manipulation benchmark MoMaRT. Code is available at https://github.com/HarryLui98/code_vpwem.

Reinforcement learning (RL) has been extensively employed in a wide range of decision-making problems, such as games and robotics. Recently, diffusion policies have shown strong potential in modeling multi-modal behaviors, enabling more diverse and flexible action generation compared to the conventional Gaussian policy. Despite various attempts to combine RL with diffusion, a key challenge is the difficulty of computing action log-likelihood under the diffusion model. This greatly hinders the direct application of diffusion policies in on-policy reinforcement learning. Most existing methods calculate or approximate the log-likelihood through the entire denoising process in the diffusion model, which can be memory- and computationally inefficient. To overcome this challenge, we propose a novel and efficient method to train a diffusion policy in an on-policy setting that requires only evaluating a simple Gaussian probability. This is achieved by aligning the policy iteration with the diffusion process, which is a distinct paradigm compared to previous work. Moreover, our formulation can naturally handle entropy regularization, which is often difficult to incorporate into diffusion policies. Experiments demonstrate that the proposed method produces multimodal policy behaviors and achieves superior performance on a variety of benchmark tasks in both IsaacLab and MuJoCo Playground.

Diffusion-based policies have achieved remarkable results in robotic manipulation but often struggle to adapt rapidly in dynamic scenarios, leading to delayed responses or task failures. We present DCDP, a Dynamic Closed-Loop Diffusion Policy framework that integrates chunk-based action generation with real-time correction. DCDP integrates a self-supervised dynamic feature encoder, cross-attention fusion, and an asymmetric action encoder-decoder to inject environmental dynamics before action execution, achieving real-time closed-loop action correction and enhancing the system's adaptability in dynamic scenarios. In dynamic PushT simulations, DCDP improves adaptability by 19% without retraining while requiring only 5% additional computation. Its modular design enables plug-and-play integration, achieving both temporal coherence and real-time responsiveness in dynamic robotic scenarios, including real-world manipulation tasks. The project page is at: https://github.com/wupengyuan/dcdp

Robotic laboratory automation has traditionally relied on carefully engineered motion pipelines and task-specific hardware interfaces, resulting in high design cost and limited flexibility. While recent imitation learning techniques can generate general robot behaviors, their large model sizes often require high-performance computational resources, limiting applicability in practical laboratory environments. In this study, we propose a compact imitation learning framework for laboratory automation using small foundation models. The proposed method, TVF-DiT, aligns a self-supervised vision foundation model with a vision-language model through a compact adapter, and integrates them with a Diffusion Transformer-based action expert. The entire model consists of fewer than 500M parameters, enabling inference on low-VRAM GPUs. Experiments on three real-world laboratory tasks - test tube cleaning, test tube arrangement, and powder transfer - demonstrate an average success rate of 86.6%, significantly outperforming alternative lightweight baselines. Furthermore, detailed task prompts improve vision-language alignment and task performance. These results indicate that small foundation models, when properly aligned and integrated with diffusion-based policy learning, can effectively support practical laboratory automation with limited computational resources.

This paper focuses on enhancing the grasping precision and generalization of manipulation policies learned via imitation learning. Diffusion-based policy learning methods have recently become the mainstream approach for robotic manipulation tasks. As grasping is a critical subtask in manipulation, the ability of imitation-learned policies to execute precise and generalizable grasps merits particular attention. Existing imitation learning techniques for grasping often suffer from imprecise grasp executions, limited spatial generalization, and poor object generalization. To address these challenges, we incorporate grasp prior knowledge into the diffusion policy framework. In particular, we employ a latent diffusion policy to guide action chunk decoding with grasp pose prior, ensuring that generated motion trajectories adhere closely to feasible grasp configurations. Furthermore, we introduce a self-supervised reconstruction objective during diffusion to embed the graspness prior: at each reverse diffusion step, we reconstruct wrist-camera images back-projected the graspness from the intermediate representations. Both simulation and real robot experiments demonstrate that our approach significantly outperforms baseline methods and exhibits strong dynamic grasping capabilities.

Multi-agent robotic manipulation remains challenging due to the combined demands of coordination, grasp stability, and collision avoidance in shared workspaces. To address these challenges, we propose the Adaptive Dynamic Modality Diffusion Policy (ADM-DP), a framework that integrates vision, tactile, and graph-based (multi-agent pose) modalities for coordinated control. ADM-DP introduces four key innovations. First, an enhanced visual encoder merges RGB and point-cloud features via Feature-wise Linear Modulation (FiLM) modulation to enrich perception. Second, a tactile-guided grasping strategy uses Force-Sensitive Resistor (FSR) feedback to detect insufficient contact and trigger corrective grasp refinement, improving grasp stability. Third, a graph-based collision encoder leverages shared tool center point (TCP) positions of multiple agents as structured kinematic context to maintain spatial awareness and reduce inter-agent interference. Fourth, an Adaptive Modality Attention Mechanism (AMAM) dynamically re-weights modalities according to task context, enabling flexible fusion. For scalability and modularity, a decoupled training paradigm is employed in which agents learn independent policies while sharing spatial information. This maintains low interdependence between agents while retaining collective awareness. Across seven multi-agent tasks, ADM-DP achieves 12-25% performance gains over state-of-the-art baselines. Ablation studies show the greatest improvements in tasks requiring multiple sensory modalities, validating our adaptive fusion strategy and demonstrating its robustness for diverse manipulation scenarios.

Effective robotic manipulation requires policies that can anticipate physical outcomes and adapt to real-world environments. Effective robotic manipulation requires policies that can anticipate physical outcomes and adapt to real-world environments. In this work, we introduce a unified framework, World-Model-Driven Diffusion Policy with Online Adaptive Learning (AdaWorldPolicy) to enhance robotic manipulation under dynamic conditions with minimal human involvement. Our core insight is that world models provide strong supervision signals, enabling online adaptive learning in dynamic environments, which can be complemented by force-torque feedback to mitigate dynamic force shifts. Our AdaWorldPolicy integrates a world model, an action expert, and a force predictor-all implemented as interconnected Flow Matching Diffusion Transformers (DiT). They are interconnected via the multi-modal self-attention layers, enabling deep feature exchange for joint learning while preserving their distinct modularity characteristics. We further propose a novel Online Adaptive Learning (AdaOL) strategy that dynamically switches between an Action Generation mode and a Future Imagination mode to drive reactive updates across all three modules. This creates a powerful closed-loop mechanism that adapts to both visual and physical domain shifts with minimal overhead. Across a suite of simulated and real-robot benchmarks, our AdaWorldPolicy achieves state-of-the-art performance, with dynamical adaptive capacity to out-of-distribution scenarios.

Realizing dexterous embodied manipulation necessitates the deep integration of heterogeneous multimodal sensory inputs. However, current vision-centric paradigms often overlook the critical force and geometric feedback essential for complex tasks. This paper presents DeMUSE, a Deep Multimodal Unified Sparse Experts framework leveraging a Diffusion Transformer to integrate RGB, depth, and 6-axis force into a unified serialized stream. Adaptive Modality-specific Normalization (AdaMN) is employed to recalibrate modality-aware features, mitigating representation imbalance and harmonizing the heterogeneous distributions of multi-sensory signals. To facilitate efficient scaling, the architecture utilizes a Sparse Mixture-of-Experts (MoE) with shared experts, increasing model capacity for physical priors while maintaining the low inference latency required for real-time control. A Joint denoising objective synchronously synthesizes environmental evolution and action sequences to ensure physical consistency. Achieving success rates of 83.2% and 72.5% in simulation and real-world trials, DeMUSE demonstrates state-of-the-art performance, validating the necessity of deep multi-sensory integration for complex physical interactions.

Imitation learning for robotic manipulation has progressed from 2D image policies to 3D representations that explicitly encode geometry. Yet purely geometric policies often lack explicit part-level semantics, which are critical for pose-aware manipulation (e.g., distinguishing a shoe's toe from heel). In this paper, we present HeRO, a diffusion-based policy that couples geometry and semantics via hierarchical semantic fields. HeRO employs dense semantics lifting to fuse discriminative, geometry-sensitive features from DINOv2 with the smooth, globally coherent correspondences from Stable Diffusion, yielding dense features that are both fine-grained and spatially consistent. These features are processed and partitioned to construct a global field and a set of local fields. A hierarchical conditioning module conditions the generative denoiser on global and local fields using permutation-invariant network architecture, thereby avoiding order-sensitive bias and producing a coherent control policy for pose-aware manipulation. In various tests, HeRO establishes a new state-of-the-art, improving success on Place Dual Shoes by 12.3% and averaging 6.5% gains across six challenging pose-aware tasks. Code is available at https://github.com/Chongyang-99/HeRO.

Online Multi-Agent Reinforcement Learning (MARL) is a prominent framework for efficient agent coordination. Crucially, enhancing policy expressiveness is pivotal for achieving superior performance. Diffusion-based generative models are well-positioned to meet this demand, having demonstrated remarkable expressiveness and multimodal representation in image generation and offline settings. Yet, their potential in online MARL remains largely under-explored. A major obstacle is that the intractable likelihoods of diffusion models impede entropy-based exploration and coordination. To tackle this challenge, we propose among the first Online off-policy MARL framework using Diffusion policies (OMAD) to orchestrate coordination. Our key innovation is a relaxed policy objective that maximizes scaled joint entropy, facilitating effective exploration without relying on tractable likelihood. Complementing this, within the centralized training with decentralized execution (CTDE) paradigm, we employ a joint distributional value function to optimize decentralized diffusion policies. It leverages tractable entropy-augmented targets to guide the simultaneous updates of diffusion policies, thereby ensuring stable coordination. Extensive evaluations on MPE and MAMuJoCo establish our method as the new state-of-the-art across $10$ diverse tasks, demonstrating a remarkable $2.5\times$ to $5\times$ improvement in sample efficiency.

Large-scale Vision-Language-Action (VLA) models offer semantic generalization but suffer from high inference latency because they adopt a low-frequency batch-and-execute paradigm. This frequency mismatch creates an execution blind spot, causing failures in dynamic environments where targets move during the open-loop execution window. We propose TIDAL (Temporally Interleaved Diffusion and Action Loop), a hierarchical framework that decouples semantic reasoning from high-frequency actuation. TIDAL operates as a backbone-agnostic scheduler for diffusion-based VLAs, using a dual-frequency architecture to redistribute the computational budget. Specifically, a low-frequency macro-intent loop caches semantic embeddings, while a high-frequency micro-control loop interleaves single-step flow integration with execution, conditioning on the latest state fused with motion cues. To handle the resulting latency shift, we introduce a temporally misaligned training strategy where the policy learns stalenessaware compensation, conditioning on stale semantic intent alongside real-time proprioception. TIDAL is architectural, making it orthogonal to system-level optimizations. Experiments show an average 2.5x performance gain over open-loop baselines in dynamic interception tasks. Despite a slight decrease in static success rates, our approach yields a 4x increase in feedback frequency and extends the effective horizon of semantic embeddings beyond the native action chunk size. Under nonpaused physics and on a real robot, TIDAL remains robust to unmodeled dynamics, while standard open-loop baselines fail due to latency-induced error accumulation.

Vision-based imitation learning has enabled impressive robotic manipulation skills, but action imitation alone provides limited supervision of the geometric consequences of robot behavior. To address this limitation, we introduce Implicit Scene Supervision (ISS) Policy, a 3D visuomotor diffusion policy with a DiT backbone that predicts continuous action sequences from point-cloud observations. ISS augments action diffusion with a supervised robot motion predictor that maps generated actions and robot-state context to end-effector motion, and then uses the predicted motion together with gripper intent to forecast future point-cloud representations. By explicitly modeling the intermediate transition from action to robot motion, ISS encourages the policy to capture how its actions affect the surrounding 3D scene. We further introduce asymmetric gradient routing to separate direct motion regression from scene-level policy supervision, together with a change-balanced objective that accounts for variations in scene-change magnitude. These auxiliary objectives provide dynamics-aware geometric supervision using only expert demonstrations, without requiring additional annotations or auxiliary modules at inference time. ISS Policy achieves state-of-the-art performance on single-arm manipulation tasks in MetaWorld and dexterous manipulation tasks in Adroit, while real-world dual-arm experiments further demonstrate its effectiveness on physical robotic manipulation. The resulting framework preserves the scalable DiT backbone and standard diffusion-policy control interface. Code and videos will be released.

Diffusion Policy (DP) excels in embodied control but suffers from high inference latency and computational cost due to multiple iterative denoising steps. Existing diffusion model (DMs) acceleration methods improve inference speed by reducing computational overhead, as DMs for image generation are computation-bound. In this work, we reveal that DP are I/O bound, and thus propose using speculative decoding, a parallel computing framework, to accelerate DP for the first time. However, it is non-trivial to address the following challenges: (1) How to preserve the denoising quality of the base model at reduced computational cost under time-varying task difficulty in embodied settings; and (2) How to dynamically allocate computation according to task difficulty in such environments. Therefore, we propose Temporal-aware Reinforcement-based Speculative Diffusion Policy (TS-DP), the first framework that enables speculative decoding for DP with temporal adaptivity. First, to handle dynamic environments where task difficulty varies over time, we distill a Transformer-based drafter to imitate the base model and replace its costly denoising calls. Second, an RL-based scheduler further adapts to the time-varying task difficulty by adjusting speculative parameters to maintain accuracy while improving efficiency. Extensive experiments show that TS-DP achieves up to a 4.17x NFE reduction and a 3.2x wall-clock speedup with a draft acceptance rate of over 85% in simulation. Real-world experiments further demonstrate a 3.8x NFE reduction while improving success rates on real-time tasks (e.g., a 23% gain on the conveyor-belt task), reaching an inference frequency of 26 Hz and enabling real-time control while maintaining comparable task performance. The code is available at https://github.com/ChildTang/TS-DP.

Equivariant diffusion policies (EDPs) combine the expressivity of diffusion models with the generalization and sample efficiency afforded by geometric symmetries. While steering these policies with reinforcement learning (RL) offers a promising fine-tuning mechanism beyond demonstration data, directly applying standard (non-equivariant) RL ignores the symmetries that EDPs are designed to exploit, leading to sample inefficiency and instability. We theoretically establish that the diffusion process of an EDP is equivariant, inducing a group-invariant latent-noise MDP which is well-suited for equivariant diffusion steering. Building on this, we introduce a principled symmetry-aware steering framework and compare standard, equivariant, and approximately equivariant RL strategies across tasks with varying degrees of symmetry. While we identify the practical boundaries of strict equivariance under symmetry breaking, exploiting symmetry during the steering yields substantial benefits---enhancing sample efficiency, preventing value divergence, and achieving strong policy improvements even when EDPs are trained from extremely limited demonstrations.

Diffusion models provide an expressive framework for sampling from complex, unnormalized distributions. In this work, we extend Maximum Entropy Reinforcement Learning (ME-RL) to diffusion-based policies by introducing Diffusion-Augmented Markov Decision Processes (DA-MDPs). DA-MDPs interpret each reverse-diffusion transition as an individual reinforcement-learning decision, while only the final denoised action is executed in the environment. Our DA-MDPs follow from a principled derivation based on the variational-inference formulation of ME-RL. By augmenting policy and target trajectories with intermediate diffusion variables, we obtain a tractable reverse-KL upper bound via the data-processing inequality. This bound decomposes across denoising transitions, yielding diffusion-augmented variants of soft rewards, value functions, and local policy objectives. This provides a general framework for adapting ME-RL algorithms to diffusion policies while differentiating through only one diffusion transition at a time. We instantiate the framework with PPO, REPPO, and a maximum-entropy extension of WPO. Experiments demonstrate improved continuous-control performance, benefits from additional diffusion steps, and memory-efficient training. On the StackCube and PushT manipulation tasks, DA-MDP methods learn alternative successful strategies from the same initial state and achieve higher success rates and generally higher success-weighted mode entropy than the Gaussian ME-RL baseline. We also demonstrate successful training when using action chunking.

Ad hoc teamwork (AHT) requires agents to collaborate with previously unseen teammates, which is crucial for many real-world applications. The core challenge of AHT is to develop an ego agent that can predict and adapt to unknown teammates on the fly. Conventional RL-based approaches optimize a single expected return, which often causes policies to collapse into a single dominant behavior, thus failing to capture the multimodal cooperation patterns inherent in AHT. In this work, we introduce PADiff, a diffusion-based approach that captures agent's multimodal behaviors, unlocking its diverse cooperation modes with teammates. However, standard diffusion models lack the ability to predict and adapt in highly non-stationary AHT scenarios. To address this limitation, we propose a novel diffusion-based policy that integrates critical predictive information about teammates into the denoising process. Extensive experiments across three cooperation environments demonstrate that PADiff outperforms existing AHT methods significantly.

As robots become more integrated in society, their ability to coordinate with other robots and humans on multi-modal tasks (those with multiple valid solutions) is crucial. Such behaviors can be learned from expert demonstrations via imitation learning (IL), but when expert demonstrations are multi-modal, standard IL approaches usually average across modes or collapse to a single mode, preventing effective coordination. Being inspired by diffusion models' ability to capture complex multi-modal trajectory distributions in single-agent settings, we develop a diffusion-based framework for coordinated multi-modal behavior in multi-agent systems. However, existing multi-agent diffusion approaches typically require a centralized planner or explicit communication among agents. This assumption can fail in real-world scenarios where robots must operate independently or with agents like humans that they cannot directly communicate with. Therefore, we propose MIMIC-D, a joint training with decentralized execution paradigm for multi-modal multi-agent IL via diffusion. We jointly train all agents' policies with only local information to achieve implicit coordination. In simulation and hardware experiments, our method exhibits robust multi-modal coordination behavior in various tasks and environments, improving upon state-of-the-art baselines.

Visuomotor policies trained via behavior cloning are vulnerable to covariate shift, where small deviations from expert trajectories can compound into failure. Common strategies to mitigate this issue involve expanding the training distribution through human-in-the-loop corrections or synthetic data augmentation. However, these approaches are often labor-intensive, rely on strong task assumptions, or compromise the quality of imitation. We introduce Latent Policy Barrier, a framework for robust visuomotor policy learning. Inspired by Control Barrier Functions, LPB treats the latent embeddings of expert demonstrations as an implicit barrier separating safe, in-distribution states from unsafe, out-of-distribution (OOD) ones. Our approach decouples the role of precise expert imitation and OOD recovery into two separate modules: a base diffusion policy solely on expert data, and a dynamics model trained on both expert and suboptimal policy rollout data. At inference time, the dynamics model predicts future latent states and optimizes them to stay within the expert distribution. Both simulated and real-world experiments show that LPB improves both policy robustness and data efficiency, enabling reliable manipulation from limited expert data and without additional human correction or annotation.