Haotian Lin

Traditional robot programming is challenging: it requires orchestrating multimodal perception, managing physical contact dynamics, and handling diverse configurations and execution failures. We introduce ASPIRE (Agentic Skill Programming through Iterative Robot Exploration), a continual learning system that autonomously writes and refines robot control programs in a code-as-policy paradigm while compounding experience into a reusable skill library. ASPIRE discovers skills that persist across tasks, simulation and real-world settings, and embodiments. It operates in an open-ended loop with three components: (1) a closed-loop robot execution engine that exposes fine-grained multimodal traces, enabling autonomous failure diagnosis, repair synthesis, and validation; (2) a continually expanding skill library that distills validated fixes into reusable, transferable knowledge; and (3) evolutionary search that generates diverse task sequences and control programs to explore beyond single-trajectory refinement. ASPIRE surpasses prior methods by up to 77% on LIBERO-Pro manipulation under perturbation, 72% on Robosuite bimanual handover, and 32% on BEHAVIOR-1K long-horizon household tasks. Its accumulated library also enables zero-shot generalization to unseen long-horizon tasks: on LIBERO-Pro Long, ASPIRE achieves 31% success versus 4% for prior methods despite their use of test-time reasoning and retries. Finally, simulation-discovered skills provide initial evidence of sim-to-real transfer, substantially reducing real-robot programming effort across different embodiments and robot APIs.

Achieving dexterous robotic manipulation in the real world relies heavily on human supervision and algorithmic engineering, which is a central bottleneck in the pursuit of general physical intelligence. Although emerging coding agents can generate code to automate algorithm search, their successes remain largely confined to digital environments. We conjecture that the missing abstraction to automate robotics research is a repeatable feedback loop for real-world policy improvement: reset the scene, execute a policy, verify the outcome, and refine the next iteration. To bridge this gap, we introduce ENPIRE, a harness framework for coding agents that instantiates this physical feedback routine with four core modules: an Environment module (EN) for automatic reset and verification, a Policy Improvement module (PI) that launches policy refinement, a Rollout module (R) to evaluate policies with single or multiple physical robots operating in parallel, and an Evolution module (E) in which coding agents analyze logs, consult literature, improve training infrastructure and algorithm code to address failure modes. This closed-loop system transforms real-world robot learning into a controllable optimization procedure that agents can manage, thus minimizing human effort while allowing fair ablations across training recipes and agent variants. Powered by ENPIRE, frontier coding agents can autonomously develop a policy to achieve a 99% success rate on challenging, dexterous manipulation tasks in the real world, such as PushT, organizing pins into a pin box, and using a cutter to cut a zip tie. Coding agents can improve policies with various PI regimes, such as heuristic learning, tool calling, behavior cloning, offline or online RL. Moreover, ENPIRE can be significantly accelerated on a robot fleet, and we propose two metrics, namely, Mean Robot Utilization (MRU) and Mean Token Utilization (MTU) to measure the efficiency of multiagent physical autoresearch. We also include simulation results in RoboCasa. Our findings suggest a practical and scalable path toward autonomously advancing robotics in the real world.

Supervised fine-tuning (SFT) has become the de facto post-training strategy for large vision-language-action (VLA) models, but its reliance on costly human demonstrations limits scalability and generalization. We propose Probe, Learn, Distill PLD, a plug-and-play framework that improves VLAs through residual reinforcement learning and distribution-aware data collection. In Stage 1 (specialist acquisition), we freeze the VLA backbone and train lightweight residual actors via off-policy RL. These specialists take over in states where the base policy fails, thereby probing failure regions of the generalist. In Stage 2 (data collection), we employ a hybrid rollout scheme that biases residual interventions toward states frequently visited by the base policy, aligning collected trajectories with the generalist’s deployment distribution while capturing recovery behaviors. In Stage 3 (fine-tuning), these curated trajectories are distilled back into the generalist with standard SFT, applicable to both flow-matching and autoregressive heads. We evaluate PLD across diverse settings: it achieves a near-saturated 99 % task success rate on the LIBERO benchmark, delivers over 50 % performance gains in SimplerEnv, and demonstrates practicality on real-world Franka arm manipulation tasks. We further provide ablations showing that residual policy probing and distribution-aware replay are key to collecting deployment-aligned data that improves VLAs’ capabilities on both seen and unseen tasks. Our results demonstrate that RL-generated, policy-aligned data can surpass teleoperation-only demonstrations, offering a scalable path toward self-improving VLA models.

Model-based reinforcement learning (MBRL) algorithms that integrate model predictive control with learned value or policy priors have shown great potential to solve complex continuous control problems. However, existing practice relies on online planning to collect high-quality data, resulting in value learning that is entirely dependent on off-policy experiences. Contrary to the belief that value learned from model-free policy iteration within this framework is sufficiently accurate and expressive, we found that severe value overestimation bias occurs, especially in high-dimensional tasks. Through both theoretical analysis and empirical evaluations, we identify that this overestimation stems from a structural policy mismatch: the divergence between the exploration policy induced by the model-based planner and the exploitation policy evaluated by the value prior. To improve value learning, we emphasize conservatism that mitigates out-of-distribution queries. The proposed method, TD-M(PC)\(^2\), addresses this by applying a soft-constrained policy update—a minimalist yet effective solution that can be seamlessly integrated into the existing plan-based MBRL pipeline without incurring additional computational overhead. Extensive experiments demonstrate that the proposed approach improves performance over baselines by large margins, particularly in 61-DoF humanoid control tasks.

Learning domain-adaptive policies that can generalize to unseen transition dynamics remains a fundamental challenge in learning-based control. Substantial progress has been made through domain representation learning to capture domain-specific information, thus enabling domain-aware decision making. We analyze the process of learning domain representations through dynamical prediction and find that selecting contexts adjacent to the current step causes the learned representations to entangle static domain information with varying dynamical properties. Such mixture can confuse the conditioned policy, thereby constraining zero-shot adaptation. To tackle the challenge, we propose DADP (Domain-Adaptive Diffusion Policy), which achieves robust adaptation through unsupervised disentanglement and domain-aware diffusion injection. First, we introduce Lagged Context Dynamical Prediction, a strategy that conditions future state estimation on a historical offset context; by increasing this temporal gap, we unsupervisedly disentangle static domain representations by filtering out transient properties. Second, we integrate the learned domain representations directly into the generative process by biasing the prior distribution and reformulating the diffusion target. Extensive experiments on challenging benchmarks across locomotion and manipulation demonstrate the superior performance and the generalizability of DADP over prior methods. More visualization results are available on the website.