Learning from Suboptimal Demonstrations

Human-collected robot manipulation data often mix successful behavior with retries, pauses, recoveries, and failures. Advantage-Conditioned Behavior Cloning (AC-BC) can exploit such mixed-quality data, but only if the conditioning signal is both accurately constructed from offline trajectories and effectively incorporated into the policy's action-vector-field prediction. Existing pipelines can weaken this signal by thresholding estimator-derived advantage scores into noisy conditions, and by routing conditions through prompts that weakly affect the action vector field. We introduce ALIGN, a condition-aligned AC-BC framework for vision-language-action policies. ALIGN directly predicts categorical pairwise conditions in the policy-facing label space and injects them into a DiT action expert through AdaRMS modulation fused with flow time. We evaluate ALIGN on deliberately mixed real-robot datasets for ARX X5 bowl stacking and Franka key hanging, each combining standard demonstrations with retry-heavy trajectories. ALIGN improves condition fidelity, strengthens conditional vector-field separability, and achieves 90.0% and 56.7% closed-loop success, compared with 56.7% and 33.3% for the strongest AC-BC baseline. These results indicate that aligning conditions from estimation to action prediction is an effective way to extract robust manipulation policies from data in which success, recovery and error coexist.

Classical inverse bandit methods typically aim to recover true reward parameters. In contrast, we infer a demonstrator's subjective beliefs, specifically its internal estimates of suboptimality gaps. We consider no-regret demonstrators that follow randomized policies, such as Thompson sampling, and characterize a family of algorithms for which per-step inversion is possible. This enables gap estimates to be recovered directly from the arm-selection distribution. Our analysis identifies an inherent stability-coverage trade-off: inversion becomes ill-conditioned as the demonstrator concentrates its probability mass on a single arm, while the number of informative rounds depends on the rate at which its policy converges. We establish conditions under which inversion is reliable and provide theoretical guarantees on the resulting estimation error. Experiments on simulated data support the analysis and illustrate how these conditions affect gap recovery in practice.

Progress estimators play an essential role in providing dense reward signals for reinforcement learning in robotic manipulation, but accurately assessing task progress can be much challenging. Existing methods heavily rely on expert data and naively assume that all actions within a demonstration contribute to task success. This assumption often breaks down in real-world scenarios, especially in high-precision manipulation, where subtle geometric errors can lead to task failure. Therefore, these methods can erroneously interpret human mistakes as meaningful progress, further corrupting progress learning. We argue that a key limitation is the lack of a global prior regarding task evaluation criteria, which is crucial for distinguishing meaningful progress from human mistakes. To this end, we propose PRIME, a novel progress estimator capable of learning precise progress estimation from imprecise demonstrations by aggregating experience across trajectories. PRIME constructs a weighted graph over observed states, connecting states across trajectories based on spatial proximity and visual consistency. Edge weights encode spatial distances as transition costs, and shortest-path search approximates the geodesic distance to success. Changes in this distance naturally distinguish advancement, stagnation, and setbacks without assuming monotonic progress over time. Extensive experiments in both simulation and real world on precise manipulation tasks show that PRIME is consistently effective across diverse tasks, and robust to data with varying degrees of suboptimality. Notably, the advantage of PRIME gets more pronounced in higher-precision scenarios, yielding up to a $5\times$ improvement in policy success rate over the SFT baselines on charger insertion.

Scaling expert demonstrations alone does not enable generalist robot policies to improve from deployment. Real-world rollouts contain successes, suboptimal behavior, and failures: imitating every recorded action can reinforce poor behavior, while filtering out failures discards evidence of physical interactions and their consequences. We introduce LoopWAM, a closed-loop world-action model that learns directly from this full quality spectrum, using rollouts from both itself and other deployed policies. Source-dependent noise sampling unifies behavior learning and action-conditioned world modeling, while quality steering and quality-weighted supervision retain full video supervision across qualities and regulate action imitation. Video and action tokens interact throughout parallel denoising, allowing learned consequences to inform action generation directly. On RoboChallenge Table30-V2, three deployment-and-training rounds raise success from 23.67% to 43.33%, a 19.7-percentage-point gain, with a final score of 56.80. LoopWAM also achieves 94.5% average success on RoboTwin 2.0 and 49.5% overall success on RoboCasa365 without additional embodied pretraining, obtaining the best overall results among the compared methods across all three benchmarks. On challenging shirt folding task, two rounds of deployment-driven improvement after single-task post-training raise success from 13.20% to 55.56%. These results demonstrate closed-loop improvement from mixed-quality experience under both multi-task learning and single-task post-training. Project page: https://loop-wam.github.io/

Robots can acquire bad habits from a few defective moments in otherwise useful teleoperation. In mixed-quality robot data, normal and problematic demonstrations share most task behavior and differ only at a local action continuation. Full retraining is costly, while fine-tuning on clean data alone offers limited recovery under a small update budget. We ask whether the local difference can instead be fixed with only 1% of the full-training sample-backward budget. Using only episode-level retained/problematic labels, REDIRECT localizes the branch, assigns a coherent retained continuation to the original observations around it, and anchors shared behavior, without frame-level annotations or additional interaction. Across three randomized ManiSkill tasks and three seeds, REDIRECT raises the mean success rate from 68.2% to 90.3%, recovering 88.8% of the clean-retraining gap versus 22.1% for matched-compute fine-tuning. On a PiPER arm, it recovers 86.7-92.9% of the clean-only gap across cup insertion and towel folding. Local robot habits can therefore be repaired by spending the update budget on the behavioral difference rather than relearning shared behavior.

Many generative systems return multiple candidate responses and are evaluated according to the best one. Recent work shows that, when demonstrations are optimal, pass@$k$ can reduce the sample complexity of learning from demonstrations by a logarithmic factor in $k$. We ask what happens when the demonstrator is not assumed to be optimal. We find that multiple responses provide a qualitatively stronger benefit in this setting. In a finite reward-class model with no reward feedback, moving from pass@$1$ to any pass@$k$ with $k\ge2$ changes the worst-case dependence on target accuracy from $1/\varepsilon^2$ to $1/\varepsilon$, uniformly over demonstrator quality. Under standard evaluation, where an unknown reward is fixed before training, increasing $k$ provides an additional and distinct benefit: the optimal dependence on a reward class of size $N$ improves from $\log N$ to $\log N/\log k$. We further show that these two effects can be separated. Under robust evaluation, where one learned policy must compete with the demonstrator simultaneously for every reward in the class, the fast $1/\varepsilon$ dependence persists, while the $1/\log k$ improvement can disappear. We establish matching upper and lower bounds in the corresponding regimes and give a greedy multiplicative-weights learner achieving the upper bounds without any assumption on demonstrator quality.

Prior work on imitation learning from suboptimal demonstrations typically relies on compressed supervision signals such as confidence estimates, discriminator scores, or importance weights. These scalar signals are inherently limited, as they cannot explicitly express intermediate reasoning about task progress, failure modes, or corrective actions. We propose a language-critique framework for imitation learning from suboptimal demonstrations that instead leverages natural language as a structured supervision signal, avoiding the collapse of expressive feedback into scalars. Our method first constructs language labels from demonstrations that explicitly describe current progress, identify suboptimal behaviors, and provide fine-grained corrective guidance. We then introduce a language-critique loss that directly trains policies using these structured signals without reducing them to scalars, and instantiate it for both behavior cloning and diffusion policies, yielding LC-BC and LC-DP. We further provide a theoretical result showing that the proposed objective upper-bounds the expert performance gap under standard assumptions. Empirically, we evaluate on diverse continuous control tasks spanning navigation, manipulation, and gameplay, where our methods consistently outperform strong imitation learning and offline reinforcement learning baselines. These results demonstrate that language can serve as a powerful and structured form of supervision for learning robust policies from suboptimal data.

Human demonstrations for robot imitation learning often contain mistakes and corrective behaviors, such as imprecise grasps, object misalignment, unstable contact, and repeated attempts. While these segments are commonly treated as noisy or suboptimal data, they provide valuable evidence about when execution deviates from a desirable path and how task feasibility can be restored. However, existing reward and value models often rely on monotonic progress assumptions, which capture coarse task advancement but may overlook local execution errors and corrective behaviors in imperfect demonstrations. In this work, we propose ReTVL (ReTry-Supervised Value Learning), a framework for learning mistake-sensitive value functions from mixed-quality robot demonstrations by leveraging retry events as sparse supervision. ReTVL captures the local degradation-and-recovery structure around mistakes by combining global progress calibration with local pairwise preference learning induced by sparsely annotated retry keypoints. The learned value model is then used to reweight demonstration chunks for downstream behavior cloning, reducing the influence of harmful execution errors while preserving useful corrective behaviors. Experiments on real-robot manipulation tasks show that ReTVL produces more fine-grained value estimates than progress-based baselines and improves imitation learning from imperfect demonstrations.

We propose Ambient Diffusion Policy, a simple and principled method for imitation learning from suboptimal data in robotics. High-quality, task-specific robot data is expensive and time-consuming to collect, while suboptimal datasets with lower-quality or out-of-distribution demonstrations are abundant. Existing methods that co-train on both data sources in robotics often fail to separate the meaningful and the harmful features in the suboptimal samples. In contrast, our method extracts only the useful features by introducing a new axis to co-training in robotics: noise-dependent data usage. Ambient Diffusion Policy restricts the contribution of suboptimal data during training to only the high and low diffusion times. To rigorously justify our approach, we first observe that robot action data exhibits a spectral power law. This induces two important properties on the optimal Diffusion Policy that we exploit: a global-to-local hierarchy and locality. We theoretically formalize this discussion using a simplified model. Our experiments validate Ambient Diffusion Policy on four types of suboptimal action data (noisy trajectories, sim-to-real gap, task mismatch, and large-scale data mixtures) across six tasks. The results show that it effectively learns from arbitrary sources of suboptimal data. Notably, it outperforms existing co-training baselines by up to 33% when scaled to Open X-Embodiment - a large dataset with heterogeneous data quality and unstructured distribution shifts. Overall, Ambient Diffusion Policy increases the utility of suboptimal demonstrations and expands the set of usable data sources in robotics.

Inverse reinforcement learning (IRL) typically assumes demonstrations from a single optimal demonstrator, but in many applications data come from multiple imperfect demonstrators with heterogeneous suboptimality levels. We study reward learning in this setting through a feasible-reward-set framework: for each demonstrator, we encode its declared suboptimality level as a linear constraint and intersect the resulting feasible sets across demonstrators. Our theoretical analysis shows that the joint feasible set shrinks monotonically as data are added, and we give an exact characterization of when a new demonstrator strictly tightens it. We further establish two recovery guarantees for the feasible reward set of the ground-truth optimal demonstrator: one bound depends on closeness to the optimal occupancy, while the other requires only sufficient coverage and no near-optimal demonstrator. On the practical side, we introduce strategies to address the inherent reward ambiguity in the obtained reward set and provide an offline algorithm with function approximation for high-dimensional environments. Experiments in tabular grid-world and large language model (LLM) fine-tuning settings are consistent with the theoretical predictions and demonstrate the effectiveness of the proposed framework over baselines.

Learning reward functions from demonstrations assumes that demonstrations provide adequate supervision over all features -- or task-relevant aspects of behavior. In practice, demonstrations are often imperfect: humans may under-emphasize certain features due to cognitive load or physical difficulty, or the training regime may fail to sufficiently cover all relevant situations. In either case, important features may be underspecified, leading to ambiguity in the learned reward function and misaligned behavior at deployment. We propose a framework that detects such underspecified features and actively solicits targeted corrective demonstrations. Our key insight is that demonstrations implicitly reveal which features are well specified: features that are consistently optimized show little variation across demonstrations, while features that are underspecified vary widely. We leverage this statistical signal to infer which features may have been insufficiently demonstrated. The robot then explains which features it is uncertain about in natural language and queries for demonstrations that explicitly address the identified gaps. We evaluate our approach in a simulated tabletop manipulation domain and in a user study with a real Franka robot. Targeted, explanation-guided queries significantly improve reward recovery compared to random querying and passive data collection, reducing ambiguity that would otherwise persist in learning from imperfect demonstrations.

Robotic imitation learning typically assumes access to optimal demonstrations, yet real-world data collection often yields suboptimal, exploratory, or even failed trajectories. Discarding such data wastes valuable information about environment dynamics and failure modes, which can instead be leveraged to improve decision-making. While 3D policies reduce reliance on high-quality demonstrations through strong spatial generalization, they still require large-scale data to achieve high task success. To address this, we propose DALI-R, a Data-Asymmetric Latent Imagination and Reranking framework for 3D robotic imitation learning from mixed-quality trajectories. It learns a Latent World Model over 3D point clouds for imagined rollouts and a Task Completion Scorer that reranks candidate action chunks, improving decision-making without additional high-quality demonstrations. We instantiate DALI-R with both diffusion and efficient flow-matching policies and evaluate it on Adroit and MetaWorld benchmarks. Across the two evaluated 3D base policies, DALI-R achieves an average $6.8$% improvement in success rate while incurring less than $0.7\times$ additional inference overhead.

Imitation learning offers a promising framework for enabling robots to acquire diverse skills from human users. However, most imitation learning algorithms assume access to high-quality demonstrations an unrealistic expectation when collecting data from non-expert users, whose demonstrations often contain inadvertent errors. Naively learning from such demonstrations can result in unsafe policy behavior, while discarding entire demonstrations due to occasional mistakes wastes valuable data, especially in low-data settings. In this work, we introduce GiB (Good-in-Bad), an algorithm that automatically identifies and discards erroneous subtasks within demonstrations while preserving high-quality subtasks. The filtered data can then be used by any policy learning algorithm to train more robust policies. GiB first trains a self-supervised model to learn latent features and assigns binary weights to label each demonstration as good or bad. It then models the latent feature distribution of high-quality segments and uses the Mahalanobis distance to detect and evaluate poor-quality subtasks. We validate GiB on the Franka robot in both simulated and real-world multi-step tasks, demonstrating improved policy performance when learning from mixed-quality human demonstrations.

While apprenticeship learning has shown promise for inducing effective pedagogical policies directly from student interactions in e-learning environments, most existing approaches rely on optimal or near-optimal expert demonstrations under a fixed reward. Real-world student interactions, however, are often inherently imperfect and evolving: students explore, make errors, revise strategies, and refine their goals as understanding develops. In this work, we argue that imperfect student demonstrations are not noise to be discarded, but structured signals-provided their relative quality is ranked. We introduce HALIDE, Hierarchical Apprenticeship Learning from Imperfect Demonstrations with Evolving Rewards, which not only leverages sub-optimal student demonstrations, but ranks them within a hierarchical learning framework. HALIDE models student behavior at multiple levels of abstraction, enabling inference of higher-level intent and strategy from suboptimal actions while explicitly capturing the temporal evolution of student reward functions. By integrating demonstration quality into hierarchical reward inference,HALIDE distinguishes transient errors from suboptimal strategies and meaningful progress toward higher-level learning goals. Our results show that HALIDE more accurately predicts student pedagogical decisions than approaches that rely on optimal trajectories, fixed rewards, or unranked imperfect demonstrations.

Programming robots to perform complex tasks is often difficult and time consuming, requiring expert knowledge and skills in robot software and sometimes hardware. Imitation learning is a method for training robots to perform tasks by leveraging human expertise through demonstrations. Typically, the assumption is that those demonstrations are performed by a single, highly competent expert. However, in many real-world applications that use user demonstrations for tasks or incorporate both user data and pretrained data, such as home robotics including assistive robots, this is unlikely to be the case. This paper presents research towards a system which can leverage suboptimal demonstrations to solve ambiguous tasks; and particularly learn from its own failures. This is a negative-feedback system which achieves significant improvement over purely positive imitation learning for ambiguous tasks, achieving a 90% improvement in success rate against a system that does not utilise negative feedback, compared to a 50% improvement in success rate when utilised on a real robot, as well as demonstrating higher efficacy, memory efficiency and time efficiency than a comparable negative feedback scheme. The novel scheme presented in this paper is validated through simulated and real-robot experiments.

Assistive robots offer agency to humans with severe motor impairments. Often, these users control high-DoF robots through low-dimensional interfaces, such as using a 1-D sip-and-puff interface to operate a 6-DoF robotic arm. This mismatch results in having access to only a subset of control dimensions at a given time, imposing unintended and artificial constraints on robot motion. As a result, interface-limited demonstrations embed suboptimal motions that reflect interface restrictions rather than user intent. To address this, we present a trajectory reconstruction algorithm that reasons about task, environment, and interface constraints to lift demonstrations into the robot's full control space. We evaluate our approach using real-world demonstrations of ADL-inspired tasks performed via a 2-D joystick and 1-D sip-and-puff control interface, teleoperating two distinct 7-DoF robotic arms. Analyses of the reconstructed demonstrations and derived control policies show that lifted trajectories are faster and more efficient than their interface-constrained counterparts while respecting user preferences.

Existing approaches to reward inference typically assume that humans provide demonstrations according to specific behavior models. However, humans often indicate their goals through a wide range of behaviors, from actions that are suboptimal due to poor planning or execution to behaviors intended to communicate goals rather than achieve them. One existing solution for inferring rewards from such behavior $\unicode{x2013}$ provided it is drawn from the same distribution at training and deployment $\unicode{x2013}$ is to construct a dataset of behavior paired with known rewards, and to learn the mapping from behavior to rewards; however, prior methods in this family face notable limitations, such as restrictions to tabular settings. Given such a dataset, we propose instead that supervised learning offers a parsimonious yet powerful solution, which we term Supervised Reward Inference (SRI). Theoretically, we prove that SRI is asymptotically Bayes-optimal under standard assumptions. Empirically, SRI achieves near-ceiling performance on a prior benchmark for reward inference from suboptimal behavior, while on Meta-World robotics tasks, it infers rewards from even arbitrarily suboptimal demonstrations as accurately as those demonstrations allow. Finally, we demonstrate our framework's universality with straightforward generalizations to action- and goal-prediction.