Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.
hub Mixed citations
What Matters in Learning from Offline Human Demonstrations for Robot Manipulation
Mixed citation behavior. Most common role is background (40%).
84
Pith papers citing it
71
external citations · Pith
Background
40% of classified citations
abstract
Imitating human demonstrations is a promising approach to endow robots with various manipulation capabilities. While recent advances have been made in imitation learning and batch (offline) reinforcement learning, a lack of open-source human datasets and reproducible learning methods make assessing the state of the field difficult. In this paper, we conduct an extensive study of six offline learning algorithms for robot manipulation on five simulated and three real-world multi-stage manipulation tasks of varying complexity, and with datasets of varying quality. Our study analyzes the most critical challenges when learning from offline human data for manipulation. Based on the study, we derive a series of lessons including the sensitivity to different algorithmic design choices, the dependence on the quality of the demonstrations, and the variability based on the stopping criteria due to the different objectives in training and evaluation. We also highlight opportunities for learning from human datasets, such as the ability to learn proficient policies on challenging, multi-stage tasks beyond the scope of current reinforcement learning methods, and the ability to easily scale to natural, real-world manipulation scenarios where only raw sensory signals are available. We have open-sourced our datasets and all algorithm implementations to facilitate future research and fair comparisons in learning from human demonstration data. Codebase, datasets, trained models, and more available at https://arise-initiative.github.io/robomimic-web/
hub tools
citation-role summary
background 10
method 5
dataset 3
baseline 2
citation-polarity summary
representative citing papers
LIBERO is a new benchmark for lifelong robot learning that evaluates transfer of declarative, procedural, and mixed knowledge across 130 manipulation tasks with provided demonstration data.
Variable-length autoregressive latent sequences, trained as variational inference with a PPO-style objective, give robot policies adaptive test-time compute and yield a reusable action tokenizer.
Ambient Diffusion Policy enables better imitation learning from suboptimal robot data by leveraging spectral properties to restrict data usage to specific diffusion times.
RLDT fine-tunes pretrained flow-matching policies for continuous control by aligning them to a max-entropy RL transport field constructed via SVGD, using expected-target estimation for stable multi-step updates.
Action-only curation metrics for imitation learning fail to detect structural defects that degrade policies, while state-aware metrics recover roughly one-third of the performance gap.
Closed-Form Diffusion Policies enable training-free imitation learning by using closed-form scores derived from demonstration data, achieving competitive benchmark performance with millisecond inference and composable editing of pre-trained policies.
MiraBench defines action-conditioned reliability via three levels (physics adherence, action-following fidelity, optimism bias detection) and applies it to 12 model configurations using a 16,000-judgment human corpus, finding visual fidelity a poor proxy for action fidelity, no reliable scale benefi
WorldVLN proposes the first autoregressive world action model for aerial vision-language navigation that predicts short-horizon latent world states, decodes them to waypoints in closed loop, and uses two-stage training with Action-aware GRPO to achieve over 12% success-rate gains on benchmarks plus零
VHYDRO is a support-safe variational hybrid filter that jointly recovers continuous latent states, discrete contact modes, and sparse port-Hamiltonian laws per regime while preventing loss of feasible transitions.
Flow map policies enable fast one-step inference for flow-based RL policies, and FMQ provides an optimal closed-form Q-guided target for offline-to-online adaptation under trust-region constraints, achieving SOTA performance.
A liveness-based Bellman operator enables conservative offline policy evaluation for manipulation tasks by encoding task progression and reducing truncation bias from finite horizons.
OmniNavBench is a unified benchmark for general-purpose navigation featuring composite multi-skill instructions, support for humanoid, quadrupedal and wheeled robots, and 1779 human teleoperated trajectories across 170 environments.
HDP3 is a pocket-scale 3D diffusion policy with a Diffusion Mixer decoder that achieves state-of-the-art visuomotor control using two-step DDIM inference and under 1% of the parameters of prior 3D diffusion policies.
ScoRe-Flow achieves decoupled mean-variance control in stochastic flow matching by deriving a closed-form score for drift modulation plus learned variance, yielding faster RL convergence and higher success rates on locomotion and manipulation benchmarks.
ReV is a referring-aware visuomotor policy using coupled diffusion heads for real-time trajectory replanning in robotic manipulation, trained solely via targeted perturbations to expert demonstrations and achieving higher success rates in simulated and real tasks.
A standardized 16-task memory benchmark and 14 controlled VLA variants show perceptual memory (especially frame sampling + AdaLN modulation) is strongest overall, while symbolic subgoals win on counting and short-horizon reasoning.
UPS framework uses conformal prediction to calibrate VLM verifiers for choosing between high-confidence action execution, natural language task queries, or policy interventions, then applies residual learning from interventions to continually improve the base policy with minimal feedback.
PhysGen uses video models to learn physics for robots, outperforming baselines by up to 13.8% on Libero and matching specialized models in real-world tasks.
DSRL steers pretrained diffusion policies for robotics by applying RL to their latent noise inputs, achieving sample-efficient real-world adaptation with only black-box access.
Shortcut models enable high-quality single or few-step sampling in diffusion models with one network and training phase by conditioning on desired step size.
ATM pre-trains models to predict trajectories of any points in videos, then uses those predictions to learn strong visuomotor policies from minimal action labels, beating baselines by 80% on 130+ tasks.
A collaborative dataset spanning 22 robots and 527 skills enables RT-X models that transfer capabilities across different robot embodiments.
VIP learns a visual embedding from human videos whose distance defines dense, smooth rewards for arbitrary goal-image robot tasks without task-specific fine-tuning.
citing papers explorer
-
Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution
Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.
-
LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning
LIBERO is a new benchmark for lifelong robot learning that evaluates transfer of declarative, procedural, and mixed knowledge across 130 manipulation tasks with provided demonstration data.
-
Latent Memory Palace: Reasoning for Control as Autoregressive Variational Inference
Variable-length autoregressive latent sequences, trained as variational inference with a PPO-style objective, give robot policies adaptive test-time compute and yield a reusable action tokenizer.
-
Ambient Diffusion Policy: Imitation Learning from Suboptimal Data in Robotics
Ambient Diffusion Policy enables better imitation learning from suboptimal robot data by leveraging spectral properties to restrict data usage to specific diffusion times.
-
Reinforcement Learning for Flow-Matching Policies with Density Transport
RLDT fine-tunes pretrained flow-matching policies for continuous control by aligning them to a max-entropy RL transport field constructed via SVGD, using expected-target estimation for stable multi-step updates.
-
Auditing Demonstration Curation Metrics: Action-Only Scorers Fail on the Structural Defects That Degrade Imitation Policies
Action-only curation metrics for imitation learning fail to detect structural defects that degrade policies, while state-aware metrics recover roughly one-third of the performance gap.
-
Training-Free Imitation Learning with Closed-Form Diffusion Policies
Closed-Form Diffusion Policies enable training-free imitation learning by using closed-form scores derived from demonstration data, achieving competitive benchmark performance with millisecond inference and composable editing of pre-trained policies.
-
MiraBench: Evaluating Action-Conditioned Reliability in Robotic World Models
MiraBench defines action-conditioned reliability via three levels (physics adherence, action-following fidelity, optimism bias detection) and applies it to 12 model configurations using a 16,000-judgment human corpus, finding visual fidelity a poor proxy for action fidelity, no reliable scale benefi
-
WorldVLN: Autoregressive World Action Model for Aerial Vision-Language Navigation
WorldVLN proposes the first autoregressive world action model for aerial vision-language navigation that predicts short-horizon latent world states, decodes them to waypoints in closed loop, and uses two-stage training with Action-aware GRPO to achieve over 12% success-rate gains on benchmarks plus零
-
Support-Safe Variational Hybrid Filtering for Contact-Mode and Sparse-Law Recovery
VHYDRO is a support-safe variational hybrid filter that jointly recovers continuous latent states, discrete contact modes, and sparse port-Hamiltonian laws per regime while preventing loss of feasible transitions.
-
Aligning Flow Map Policies with Optimal Q-Guidance
Flow map policies enable fast one-step inference for flow-based RL policies, and FMQ provides an optimal closed-form Q-guided target for offline-to-online adaptation under trust-region constraints, achieving SOTA performance.
-
Offline Policy Evaluation for Manipulation Policies via Discounted Liveness Formulation
A liveness-based Bellman operator enables conservative offline policy evaluation for manipulation tasks by encoding task progression and reducing truncation bias from finite horizons.
-
Beyond Isolation: A Unified Benchmark for General-Purpose Navigation
OmniNavBench is a unified benchmark for general-purpose navigation featuring composite multi-skill instructions, support for humanoid, quadrupedal and wheeled robots, and 1779 human teleoperated trajectories across 170 environments.
-
Hyper-DP3: Frequency-Aware Right-Sizing of 3D Diffusion Policies for Visuomotor Control
HDP3 is a pocket-scale 3D diffusion policy with a Diffusion Mixer decoder that achieves state-of-the-art visuomotor control using two-step DDIM inference and under 1% of the parameters of prior 3D diffusion policies.
-
ScoRe-Flow: Complete Distributional Control via Score-Based Reinforcement Learning for Flow Matching
ScoRe-Flow achieves decoupled mean-variance control in stochastic flow matching by deriving a closed-form score for drift modulation plus learned variance, yielding faster RL convergence and higher success rates on locomotion and manipulation benchmarks.
-
Referring-Aware Visuomotor Policy Learning for Closed-Loop Manipulation
ReV is a referring-aware visuomotor policy using coupled diffusion heads for real-time trajectory replanning in robotic manipulation, trained solely via targeted perturbations to expert demonstrations and achieving higher success rates in simulated and real tasks.
-
RoboMME: Benchmarking and Understanding Memory for Robotic Generalist Policies
A standardized 16-task memory benchmark and 14 controlled VLA variants show perceptual memory (especially frame sampling + AdaLN modulation) is strongest overall, while symbolic subgoals win on counting and short-horizon reasoning.
-
When to Act, Ask, or Learn: Uncertainty-Aware Policy Steering
UPS framework uses conformal prediction to calibrate VLM verifiers for choosing between high-confidence action execution, natural language task queries, or policy interventions, then applies residual learning from interventions to continually improve the base policy with minimal feedback.
-
Learning Physics from Pretrained Video Models: A Multimodal Continuous and Sequential World Interaction Models for Robotic Manipulation
PhysGen uses video models to learn physics for robots, outperforming baselines by up to 13.8% on Libero and matching specialized models in real-world tasks.
-
Steering Your Diffusion Policy with Latent Space Reinforcement Learning
DSRL steers pretrained diffusion policies for robotics by applying RL to their latent noise inputs, achieving sample-efficient real-world adaptation with only black-box access.
-
One Step Diffusion via Shortcut Models
Shortcut models enable high-quality single or few-step sampling in diffusion models with one network and training phase by conditioning on desired step size.
-
Any-point Trajectory Modeling for Policy Learning
ATM pre-trains models to predict trajectories of any points in videos, then uses those predictions to learn strong visuomotor policies from minimal action labels, beating baselines by 80% on 130+ tasks.
-
Open X-Embodiment: Robotic Learning Datasets and RT-X Models
A collaborative dataset spanning 22 robots and 527 skills enables RT-X models that transfer capabilities across different robot embodiments.
-
VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training
VIP learns a visual embedding from human videos whose distance defines dense, smooth rewards for arbitrary goal-image robot tasks without task-specific fine-tuning.
-
DexVerse: A Modular Benchmark for Multi-Task, Multi-Embodiment Dexterous Manipulation
A modular benchmark of 100 dexterous manipulation tasks across 3 arms and 6 hands with 3,180 demonstrations reveals that current policies (Diffusion Policy, DP3, OpenVLA, π0.5) achieve only 34% mean success, exposing unsolved challenges in contact-rich and precise manipulation.
-
Optimal Transport Q-Learning for Flow Policy Steering and Acceleration
Advantage-weighted conditional optimal transport flow matching simultaneously steers flow policies toward high-value actions and straightens their integration paths, enabling 2-3 step inference while improving task success.
-
Freeform Preference Learning for Robotic Manipulation
FPL trains a language-conditioned reward model from per-axis human preferences and a reward-conditioned policy, reporting 38-point average success gains over sparse-reward and binary-preference baselines on six manipulation tasks.
-
Chronos: A Physics-Informed Full-History Framework for Non-Markovian Long-Horizon Manipulation
Chronos elevates full observation history to the policy's latent state via selective SSM tokens and a Schrödinger-inspired acceleration bridge, achieving large gains on memory-dependent robot tasks with fewer parameters.
-
Trust Your Instincts: Confidence-Driven Test-Time RL for Vision-Language-Action Models
T^2VLA is a test-time reinforcement learning framework for VLAs that uses internal confidence to define intrinsic rewards via similarity to high-confidence expert demonstrations and a dual-expert bootstrapping mechanism.
-
Keypose Exploration: Efficient Automatic Trajectory Labelling and Cross-Embodiment Policy Transfer
An automatic single-demo VLM trajectory labelling pipeline enables keypose-guided diffusion policies that match baseline performance and show preliminary benefits for cross-embodiment transfer on robomimic tasks.
-
SimFoundry: Modular and Automated Scene Generation for Policy Learning and Evaluation
An automated real-to-sim pipeline builds digital twins and affordance-preserving cousins from video, yielding sim evaluations that correlate with real robot policy success and zero-shot sim-to-real gains.
-
Learning Process Rewards via Success Visitation Matching for Efficient RL
Success Visitation Matching uses a discriminator to turn sparse outcome rewards into dense process rewards by matching visitations of successful episodes, provably preserving the optimal policy and speeding up robotic RL finetuning.
-
PoLAR: Factorizing Extent and Mode in Latent Actions for Robot Policy Learning
PoLAR imposes radial structure on latent actions in hyperbolic space to factorize extent and mode, improving robot policy performance over baselines.
-
Inductive Generalization for Robotic Manipulation
The paper introduces an inductive generalization evaluation protocol for manipulation policies and shows that SOTA vision-language-action models fail on progressively harder task variants.
-
TimeProVe: Propose, then Verify for Efficient Long Video Temporal Reasoning in Activities of Daily Living
TimeProVe proposes a propose-then-verify framework using lightweight action-based candidate evidence generation followed by targeted VLM verification for efficient long video temporal reasoning, achieving 7.3% improvement on OTB with 75% fewer VLM calls.
-
One Demo is Worth a Thousand Trajectories: Action-View Augmentation for Visuomotor Policies
A framework augments single fisheye demonstrations into multiple novel-view trajectories with obstacles via fisheye-adapted Gaussian Splatting and trajectory optimization, raising policy success rates in original and modified scenes.
-
InDex: Empowering VLA Models with Intent-Conditioned Arm-Hand Coordination for Dexterous Manipulation
InDex adapts parallel-gripper VLA policies to dexterous hands by conditioning a diffusion hand policy on a normalized scalar grasp intent, raising average simulated task success from 50.3% to 85.8%.
-
Double Preconditioning (DoPr): Optimization for Test-Time Performance, not Validation Loss
Double preconditioning (DoPr) improves downstream task performance in test-time feedback settings without consistent gains in validation loss.
-
FlowPRO: Reward-Free Reinforced Fine-Tuning of Flow-Matching VLAs via Proximalized Preference Optimization
FlowPRO applies proximalized preference optimization to flow-matching VLAs with intervention-rollback data to reach higher success rates on long-horizon bimanual tasks without rewards or critics.
-
HITL-D: Human In The Loop Diffusion Assisted Shared Control
HITL-D combines diffusion policies with human input for shared robotic control, reducing required joystick axes and improving speed and workload in manipulation tasks per a 12-participant study.
-
COBALT: Crowdsourcing Robot Learning via Cloud-Based Teleoperation with Smartphones
COBALT enables scalable crowdsourced teleoperation of robots using smartphones, supporting concurrent users with low latency and yielding a 7500+ demonstration dataset validated on imitation learning tasks.
-
DexJoCo: A Benchmark and Toolkit for Task-Oriented Dexterous Manipulation on MuJoCo
DexJoCo is a benchmark and toolkit with 11 functionally grounded tasks, 1.1K trajectories, and empirical benchmarks for task-oriented dexterous manipulation on MuJoCo.
-
SID: Sliding into Distribution for Robust Few-Demonstration Manipulation
SID achieves approximately 90% success on six real-world manipulation tasks with only two demonstrations under out-of-distribution initializations, with less than 10% performance drop under distractors and disturbances.
-
DexSynRefine: Synthesizing and Refining Human-Object Interaction Motion for Physically Feasible Dexterous Robot Actions
DexSynRefine couples HOI motion manifold flow primitives with task-space residual RL and proprioceptive adaptation to convert human-object interaction data into executable dexterous robot motions, reporting 50-70 point real-world success rate gains over kinematic retargeting on five tasks.
-
Long-Horizon Q-Learning: Accurate Value Learning via n-Step Inequalities
LQL turns n-step action-sequence lower bounds into a practical hinge-loss stabilizer for off-policy Q-learning without extra networks or forward passes.
-
When Life Gives You BC, Make Q-functions: Extracting Q-values from Behavior Cloning for On-Robot Reinforcement Learning
Q2RL extracts Q-values from a BC policy and applies Q-gating to enable efficient offline-to-online RL, outperforming baselines on D4RL/robomimic tasks and achieving up to 100% success on real-robot manipulation in 1-2 hours.
-
Learning Reactive Dexterous Grasping via Hierarchical Task-Space RL Planning and Joint-Space QP Control
A multi-agent RL high-level planner outputs task-space velocities that a GPU-parallel QP low-level controller converts to joint velocities while enforcing limits and collisions, yielding robust sim-to-real dexterous grasping with zero-shot steerability.
-
OGPO: Sample Efficient Full-Finetuning of Generative Control Policies
OGPO enables sample-efficient full-finetuning of generative control policies via off-policy critics and modified PPO, achieving SOTA on robot manipulation tasks while rescuing poorly initialized behavior cloning policies without expert data.
-
An Efficient Metric for Data Quality Measurement in Imitation Learning
Power spectral density of trajectories ranks demonstration quality for imitation learning, enabling rollout-free curation that improves fine-tuned policy success.
-
Learning from the Best: Smoothness-Driven Metrics for Data Quality in Imitation Learning
RINSE scores robot demonstration trajectories for smoothness via SAL and TED metrics to curate higher-quality data for behavioral cloning, improving success rates with less data on benchmarks and real robots.