World Action Models Enable Continual Imitation Learning with Recurrent Generative Replays

Going beyond predicting robot actions, World Action Models (WAMs) can also generate future visual observations. We build on this generative capability to propose Recurrent Generative Replay (REGEN), a continual imitation learning framework that synthesizes pseudo-replay trajectories, enabling a robot policy to rehearse previously learned tasks without storing their original human demonstrations. During continual adaptation, REGEN recursively queries the WAM to synthesize pseudo-replay trajectories conditioned only on prior task instructions and current-task observations. Experiments in both simulation and real-world manipulation settings show that REGEN reduces catastrophic forgetting by up to $50\%$ relative to sequential fine-tuning, while approaching the performance of privileged experience replay methods that require access to real replay data. Finally, we analyze the factors limiting generated replay, identifying long-horizon visual degradation and action-observation inconsistency as the primary bottlenecks. Our results establish WAMs as a promising foundation for continual robot learning without stored demonstrations.

G2DP: Diffusion Planning with Spatio-Temporal Grid Guidance

In autonomous driving, diffusion-based planners have emerged as a promising paradigm for robust motion planning in dense and interactive traffic, as they can effectively model diverse driving behaviors. However, their inherent stochasticity often requires explicit guidance during denoising to ensure safety and route adherence for robust closed-loop execution. Existing guidance typically relies on sparse, entity-centric geometric queries or post-hoc refinement, yielding limited situational awareness and fragile performance in interactive scenes. To address this issue, we propose G2DP (Grid-Guided Diffusion Planning), a diffusion-based planner that directly enforces dense environmental constraints through inference-time guidance. Specifically, G2DP constructs a differentiable spatio-temporal cost volume by fusing probabilistic future occupancy distributions with a route-progress map. By formulating this volume as a continuous safety energy functional, it injects dense gradients directly into the denoising loop, actively steering trajectory generation toward collision-free and progress-optimal regions. Extensive closed-loop evaluations show that G2DP achieves state-of-the-art performance on nuPlan, outperforming the strongest imitation-learning baseline by +7.2 points in reactive score. It further maintains top scores in zero-shot transfers to interPlan and DeepScenario benchmarks, with collision avoidance improving by +10.15 over the unguided approach on interPlan. These results demonstrate that spatio-temporal cost grids serve as an effective representation for robust guidance in diffusion-based planning.

Ordinal Neural Collapse as a Representation Prior for Visual Navigation

Learning robust navigation policies directly from visual observations remains a fundamental challenge in vision-based robotic navigation. In end-to-end imitation learning approaches, the visual encoder and action decoder are jointly optimized using a single action loss, which provides only an indirect supervisory signal to the encoder. This indirect supervision frequently results in the encoder learning ambiguous, action-agnostic representations. The problem is further complicated by substantial variations in scene structure and appearance across diverse environments, as well as the prevalence of visual distractors inherent to real-world navigation settings. Such action-agnostic features cause the navigation policy to produce inconsistent actions at ambiguous decision points, leading to navigation failure. To overcome these limitations, we propose ORION (Ordinal Neural Collapse for Visual Navigation), a method that explicitly organizes the encoder's representation space according to the ordinal structure of navigation actions. In the context of goal-directed navigation, ego-centric control categories from Far Left to Far Right exhibit a natural ordinal relationship in which neighboring classes share similar visual contexts, while semantically opposing classes differ substantially in appearance. We encourage class representations to be arranged sequentially along a single discriminative axis, while suppressing off-axis variance within each class. The pretrained encoder is then integrated into a diffusion-based navigation framework, and the full pipeline is fine-tuned end-to-end. Extensive experiments in both simulation and real-world settings show that ORION consistently outperforms end-to-end and neural collapse baselines in navigation success rate and goal progress, with notable gains in visually challenging scenarios such as complex multi-way intersections.

VoiceTTA: Enhancing Zero-Shot Text-to-Speech via Reinforcement Learning-Based Test-Time Adaptation

Recently, zero-shot text-to-speech (TTS) has enabled high-fidelity and expressive speech synthesis, but it often fails to imitate unseen speaking styles from uncommon scenarios (e.g., crosstalk, dialects). Moreover, fine-tuning pretrained models requires large, high-quality datasets, limiting rapid personalization. We propose VoiceTTA, a reinforcement learning-based test-time adaptation (TTA) method that improves voice imitation of pretrained zero-shot TTS models. VoiceTTA introduces two style rewards based on coefficient-of-variation differences of F0 and energy, combined with speaker similarity and intelligibility (WER from a pretrained Whisper model), and optimizes learnable prefixes via group relative preference optimization (GRPO) in a flow matching-based model at inference time. Extensive experiments demonstrate substantial improvements on uncommon speech prompts, outperforming state-of-the-art baselines. Audio samples are available at https://voicetta.pages.dev/

Decoupling Semantics and Geometric Grounding: Spatial Visual Prompts for Language-Conditioned Imitation Learning

While end-to-end Vision-Language-Action (VLA) models show promise in robotic manipulation, their monolithic paradigm inherently couples semantic reasoning and spatial control. This creates a severe alignment bottleneck, limiting precise target disambiguation in data-constrained imitation learning. To overcome this, we propose SVP-IL, a decoupled architecture that explicitly extracts spatial visual grounding from the action generation loop. By leveraging vision-language foundation models, we parse instructions into zero-shot geometric masks, translating language into explicit Spatial Visual Prompts (SVP). These priors are injected into a continuous action generator via a lightweight direct feature-level fusion mechanism. This integration provides explicit and uncorrupted spatial gradient guidance while ensuring highly stable optimization under low-data regimes. Extensive experiments demonstrate that SVP-IL significantly outperforms state-of-the-art VLAs and pure visuomotor baselines. Trained on as few as 50 to 100 demonstrations, SVP-IL improves average success rates on highly ambiguous language-conditioned tasks from 24.0% to 39.5%, achieving 67.8% on standard benchmarks. Real-world robotic experiments further validate its robustness and data efficiency in unstructured physical environments.

Humanoid-DART: Humanoid Loco-Manipulation using Diffusion-guided Augmentation through Relabeling and Tracking

Imitating human demonstrations has emerged as a dominant paradigm for learning humanoid loco-manipulation policies. However, scaling these approaches remains challenging due to the high cost of collecting diverse demonstrations and the need for continual human intervention to correct policy failures. In this paper, we present a self-supervised framework that bootstraps from sparse demonstrations and progressively expands its behavioral repertoire, enabling the learning of a goal-conditioned policy that automatically explores the goal space with minimal expert supervision. Our approach combines diffusion-based trajectory generation with reinforcement learning, where the latter is used to track goal-conditioned trajectories produced by the diffusion model for a range of loco-manipulation skills. Through extensive ablation studies and comparisons with state-of-the-art methods, we demonstrate the effectiveness of our framework on multiple humanoid loco-manipulation skills.

MPC-Injection: Biasing Off-Policy Locomotion RL Toward Controller-Induced Behavior Basins

Reinforcement learning (RL) for locomotion frequently converges to locally optimal but undeployable behaviors, such as vibrating limbs or scooting on the torso, that maximize return without producing a usable gait. We present MPC-Injection, a low-overhead method that steers RL toward a designer-preferred gait by inserting transitions into the replay buffer from a model predictive controller solving the same Markov decision process. Unlike reward shaping, MPC-Injection does not require redesigning the task reward, and unlike adversarial imitation learning, it adds no discriminator, no kinematic retargeting, and no auxiliary objective. Instead, the controller's preferred behavior is transferred to the policy purely through the replay state distribution. On a 2D walker in simulation and with sim-to-real evaluation on a Go2 quadruped, we show that MPC-Injection drives the policy into the controller's behavior basin using a one to two-term task reward, producing gaits qualitatively comparable to those of reward shaping with twenty-one tuned terms and of adversarial motion priors without their discriminator and retargeting overhead. We further analyze how the injected transitions bias actor-critic updates toward controller-visited states, allowing the policy to learn behaviors that pure RL may fail to reach under simple reward functions.

Play2Perfect: What Matters in Dexterous Play Pretraining for Precise Assembly?

Multi-fingered robots promise the speed and dexterity of human hands, yet challenging problems such as precise assembly have remained out of reach. These tasks are contact-rich, making data collection for imitation learning difficult, and sparse-reward, making direct exploration with reinforcement learning (RL) intractable. Consequently, prior work has made progress by structuring the problem with specialized grippers, tool attachments, and environment fixtures. In this work, we argue that before a robot can perfect precise assembly, it must first learn to play. We further ask the question: what factors in the process of learning to play matter for precise assembly? We propose Play2Perfect, an RL framework for task-agnostic pretraining through play on diverse objects and goals, which is then perfected on precise assembly. The goal of play is to acquire reusable manipulation priors, such as grasping, in-hand reorientation and pose reaching. Finetuning then adapts this general prior to assembly, focusing exploration on the final contact-rich, high-precision interactions needed for success. We systematically study key design choices in play pretraining, including object diversity, training objective, trajectory diversity, and goal precision. We show that our prior is 33x more sample-efficient than RL training from scratch, even when provided with dense, multi-stage rewards. We demonstrate zero-shot sim-to-real transfer, achieving 60% success on tight insertions with only 0.5 mm contact clearance, and over 50% success on long-horizon multi-part assembly and screwing.

Stage-Aware and Roughness-Constrained Diffusion Policy for Multi-Stage Robotic Polishing

Polishing is a critical finishing process in high-end manufacturing fields such as aerospace, where surface quality directly affects the service performance and reliability of components. Robotic imitation learning provides a flexible solution for such tasks, but current methods remain limited in industrial polishing because of long-horizon dependencies, uncertain stage transitions, and the difficulty of modeling and regulating coupled process parameters. To address these issues, this paper proposes a Stage-Aware and Roughness-Constrained Diffusion Policy (SRDP) for robotic polishing. SRDP infers the process-stage posterior from multimodal observation histories and uses it to condition the shared reverse denoising process, enabling stage-consistent action generation without external stage labels during execution. Furthermore, a roughness-oriented process-constrained diffusion sampling method is incorporated to generate constrained feed speed and normal contact force under stage-wise preset spindle speeds, thereby improving process consistency and physical feasibility. Systematic experiments are conducted on two representative scenarios, namely spacecraft cabin coating-surface polishing and inner-cavity structural surface finishing. Comparisons with advanced baselines, ablation studies, and real-robot validations comprehensively evaluate the proposed method. The results show that SRD improves stage-transition stability, process-parameter consistency, and final surface quality across different polishing scenarios.

Scaling Nonlinear Optimization: Many Problems One GPU

Many robotics problems, including trajectory optimization, inverse kinematics, and contact-rich motion planning, reduce to nonlinear programs (NLPs). Mature NLP solvers such as IPOPT can solve these problems, offering hard constraint satisfaction, optimality guarantees, and favorable scaling with problem dimension. These solvers underpin gradient-based methods in robotics, yet remain CPU-bound and solve only one problem at a time, preventing their integration into GPU-batched learning pipelines. On the other hand, sampling-based approaches such as reinforcement learning, model predictive path integral, and imitation learning have become the core of modern robotics research due to their ability to leverage GPU-batched simulators. These simulators can generate orders of magnitude more dynamics rollouts per second than was previously possible. If a GPU-batched NLP solver existed, it would unlock similar speedups in the number of constrained, locally optimal solutions generated per second. This regime of solving many problems concurrently versus solving a single problem at a time is a key requirement for integrating NLP solvers in modern GPU-batched robotics frameworks. To this end, we introduce \texttt{jaxipm}, the first GPU-batched NLP solver, based on IPOPT, and implemented in JAX. We accomplish this by redesigning IPOPT's algorithm to eliminate control flow with \textit{heterogeneous iteration fusion}, and by minimizing GPU idle time with \textit{iteration level batching}. We evaluate \texttt{jaxipm} on a variety of quadrotor nonlinear model predictive control benchmarks, including reference tracking in the presence of obstacles, multi-quadrotor navigation without collision, and navigation in a cluttered environment. We demonstrate up to a $32.85\times$ increase in throughput over IPOPT. Our complete open-source codebase is available at https://github.com/johnviljoen/jaxipm.