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.
hub
Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware
27 Pith papers cite this work, alongside 448 external citations. Polarity classification is still indexing.
27
Pith papers citing it
448
external citations · OpenAlex
hub tools
citation-role summary
background 1
method 1
citation-polarity summary
representative citing papers
SkiP introduces action relabeling and Motion Spectrum Keying to skip redundant steps in robot trajectories, cutting executed steps by 15-40% while maintaining success rates across 72 simulated and 3 real tasks.
Betting mechanisms can yield provably more accurate and efficient estimates of real-world robot behavior than Monte Carlo sampling under specified conditions, with practical approximations demonstrated on synthetic data and a robotic manipulator task.
HTT learns shared representations across heterogeneous tactile sensors using a new paired dataset and pretraining objectives, enabling transfer to unseen sensors and tasks.
Reflective VLA improves VLA generalization on LIBERO-Plus and LIBERO-Plus-Hard by 5.4 and 4.2 percentage points by conditioning on action consequences instead of reactive single-frame inputs.
A VLA policy using view-selective visual routing and interaction-aware action MoE improves average success by 27.7% in simulation and 43.3% in real-world bimanual tasks over monolithic baselines.
Mapping point clouds to Fourier features improves high-precision imitation learning policies on RoboCasa, ManiSkill3, and real-robot tasks compared with Cartesian inputs.
Adding recurrent memory tokens to VLA models raises success rates on partially observable manipulation tasks from 0.42 to 0.84 on training and 0.07 to 0.23 on held-out tasks while preserving performance under full observability.
Dexterity-BEV creates 3D vertex-based inputs and BEV-aligned outputs to reduce spatial-temporal misalignments in end-to-end robot policies trained on diverse datasets and embodiments.
The Inverter framework formalizes inverse learning to generate coherent multi-step trajectories, outperforming offline RL and diffusion baselines on D4RL maze tasks by 24% on average with 10-100x less inference time while also matching GRAPE fidelity on single-qubit gates at >1000x speed.
SafePBDS uses pullback control barrier functions and a task manifold action interface to generate certifiably safe, steerable motions on high-DOF robots from objectives defined on arbitrary geometric spaces.
RoHIL adapts human-in-the-loop RL policies to new illumination conditions offline by combining world-model image relighting, illumination-retention replay, and anchored Bellman regularisation, improving shifted-light performance while preserving source performance on four real-robot tasks.
R&B-EnCoRe uses self-supervised importance-weighted variational inference to distill action-predictive reasoning datasets that improve VLA performance on manipulation, navigation, and driving tasks without external verifiers.
Supervised MoE on top of ACT achieves higher success in bowel grasping/retraction from <150 demos than standard ACT or generalist VLAs, with OOD robustness, unseen viewpoint generalization, and zero-shot ex vivo porcine transfer.
GraspVLA shows that pretraining a grasping model on a billion synthetic action frames enables zero-shot open-vocabulary performance and sim-to-real transfer.
Bimanual VLA coordination strategies, training recipes, and continuous action chunking transfer to unmanned aerial systems; the survey maps 183 works and lists fourteen shared research directions.
RoboLineage introduces an agent-native data lifecycle governance system that represents robot policy iteration steps as typed lineage artifacts to improve speed and auditability in real-robot workflows.
DUET pretrains collaborative policies on human-human VR demonstrations then fine-tunes on minimal robot teleoperation data, achieving equal or better performance than robot-only baselines with 5.4x faster collection across four tasks.
Autoregressive VLA policies achieve real-time execution via tokenization horizon adjustment and constrained decoding, outperforming flow-matching policies in speed and performance across simulated and real environments.
World Pilot augments VLA policies with world-action priors through latent and action steering pathways, reporting 84.7% success on LIBERO-Plus zero-shot OOD and top real-robot results across four tasks.
MilliVid compresses video frames into multi-scale token hierarchies and uses coarse-to-fine rollout in a diffusion model to maintain long-range geometric and object consistency on Minecraft videos.
PaCo-VLA adds an independent passivity shield to VLA outputs so that semantic proposals for compliance and admittance can be used in contact-rich tasks without violating passivity or causing damage.
StableVLA adds an Information Bottleneck Adapter to VLA models that improves robustness to visual corruptions by 30% on average with under 10M extra parameters and no extra data, even when using a much smaller backbone.
FAR combines failure-contrastive preference adaptation with action perturbations for test-time recovery and continual policy improvement, reporting 17.6% and 11.7% success gains over diffusion policies in simulation and real-world manipulation tasks.
citing papers explorer
-
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.
-
SkiP: When to Skip and When to Refine for Efficient Robot Manipulation
SkiP introduces action relabeling and Motion Spectrum Keying to skip redundant steps in robot trajectories, cutting executed steps by 15-40% while maintaining success rates across 72 simulated and 3 real tasks.
-
Betting for Sim-to-Real Performance Evaluation
Betting mechanisms can yield provably more accurate and efficient estimates of real-world robot behavior than Monte Carlo sampling under specified conditions, with practical approximations demonstrated on synthetic data and a robotic manipulator task.
-
Heterogeneous Tactile Transformer
HTT learns shared representations across heterogeneous tactile sensors using a new paired dataset and pretraining objectives, enabling transfer to unseen sensors and tasks.
-
Reflective VLA: In-Context Action Consequences Make VLAs Generalize
Reflective VLA improves VLA generalization on LIBERO-Plus and LIBERO-Plus-Hard by 5.4 and 4.2 percentage points by conditioning on action consequences instead of reactive single-frame inputs.
-
See Selectively, Act Adaptively: Dual-Level Structural Decomposition for Bimanual Robot Manipulation
A VLA policy using view-selective visual routing and interaction-aware action MoE improves average success by 27.7% in simulation and 43.3% in real-world bimanual tasks over monolithic baselines.
-
Fourier Features Let Agents Learn High Precision Policies with Imitation Learning
Mapping point clouds to Fourier features improves high-precision imitation learning policies on RoboCasa, ManiSkill3, and real-robot tasks compared with Cartesian inputs.
-
$\mu$VLA: On Recurrent Memory for Partially Observable Manipulation in VLA Models
Adding recurrent memory tokens to VLA models raises success rates on partially observable manipulation tasks from 0.42 to 0.84 on training and 0.07 to 0.23 on held-out tasks while preserving performance under full observability.
-
Dexterity-BEV: Aligning 3D World and Actions for Generalizable Robot Policies Learning
Dexterity-BEV creates 3D vertex-based inputs and BEV-aligned outputs to reduce spatial-temporal misalignments in end-to-end robot policies trained on diverse datasets and embodiments.
-
Neuro-Inspired Inverse Learning for Planning and Control
The Inverter framework formalizes inverse learning to generate coherent multi-step trajectories, outperforming offline RL and diffusion baselines on D4RL maze tasks by 24% on average with 10-100x less inference time while also matching GRAPE fidelity on single-qubit gates at >1000x speed.
-
Safe and Steerable Geometric Motion Policies for Robotic Dexterous Manipulation
SafePBDS uses pullback control barrier functions and a task manifold action interface to generate certifiably safe, steerable motions on high-DOF robots from objectives defined on arbitrary geometric spaces.
-
RoHIL: Robust Human-in-the-Loop Robotic Reinforcement Learning Against Illumination Variations
RoHIL adapts human-in-the-loop RL policies to new illumination conditions offline by combining world-model image relighting, illumination-retention replay, and anchored Bellman regularisation, improving shifted-light performance while preserving source performance on four real-robot tasks.
-
Self-Supervised Bootstrapping of Action-Predictive Embodied Reasoning
R&B-EnCoRe uses self-supervised importance-weighted variational inference to distill action-predictive reasoning datasets that improve VLA performance on manipulation, navigation, and driving tasks without external verifiers.
-
Supervised Mixture-of-Experts for Surgical Grasping and Retraction
Supervised MoE on top of ACT achieves higher success in bowel grasping/retraction from <150 demos than standard ACT or generalist VLAs, with OOD robustness, unseen viewpoint generalization, and zero-shot ex vivo porcine transfer.
-
GraspVLA: a Grasping Foundation Model Pre-trained on Billion-scale Synthetic Action Data
GraspVLA shows that pretraining a grasping model on a billion synthetic action frames enables zero-shot open-vocabulary performance and sim-to-real transfer.
-
Vision Language Action (VLA) Models for Unmanned Aerial Robotics and Bimanual Manipulation: A Review
Bimanual VLA coordination strategies, training recipes, and continuous action chunking transfer to unmanned aerial systems; the survey maps 183 works and lists fourteen shared research directions.
-
RoboLineage: Agent-Native Data Lifecycle Governance Across Robot Policy Iterations
RoboLineage introduces an agent-native data lifecycle governance system that represents robot policy iteration steps as typed lineage artifacts to improve speed and auditability in real-robot workflows.
-
Duet: Dual-Robot Understanding via Efficient Teaching
DUET pretrains collaborative policies on human-human VR demonstrations then fine-tunes on minimal robot teleoperation data, achieving equal or better performance than robot-only baselines with 5.4x faster collection across four tasks.
-
Real-Time Execution with Autoregressive Policies
Autoregressive VLA policies achieve real-time execution via tokenization horizon adjustment and constrained decoding, outperforming flow-matching policies in speed and performance across simulated and real environments.
-
World Pilot: Steering Vision-Language-Action Models with World-Action Priors
World Pilot augments VLA policies with world-action priors through latent and action steering pathways, reporting 84.7% success on LIBERO-Plus zero-shot OOD and top real-robot results across four tasks.
-
MilliVid: Hierarchical Latents for Long-Range Consistency in Video Generation
MilliVid compresses video frames into multi-scale token hierarchies and uses coarse-to-fine rollout in a diffusion model to maintain long-range geometric and object consistency on Minecraft videos.
-
PaCo-VLA: Passivity-Shielded Compliance Prior for Contact-Rich Vision-Language-Action Manipulation
PaCo-VLA adds an independent passivity shield to VLA outputs so that semantic proposals for compliance and admittance can be used in contact-rich tasks without violating passivity or causing damage.
-
StableVLA: Towards Robust Vision-Language-Action Models without Extra Data
StableVLA adds an Information Bottleneck Adapter to VLA models that improves robustness to visual corruptions by 30% on average with under 10M extra parameters and no extra data, even when using a much smaller backbone.
-
FAR: Failure-Aware Retry for Test-Time Recovery and Continual Policy Improvement
FAR combines failure-contrastive preference adaptation with action perturbations for test-time recovery and continual policy improvement, reporting 17.6% and 11.7% success gains over diffusion policies in simulation and real-world manipulation tasks.
-
StereoPolicy: Improving Robotic Manipulation Policies via Stereo Perception
StereoPolicy fuses left-right image features via cross-attention to deliver consistent gains over RGB, RGB-D, point cloud, and multi-view baselines in simulation and real-robot manipulation tasks.
-
EgoLive: A Large-Scale Egocentric Dataset from Real-World Human Tasks
EgoLive is presented as the largest open-source annotated egocentric dataset for real-world task-oriented human routines, captured with a custom head-mounted device and multi-modal annotations exclusively in unconstrained environments.
- FACTR 2: Learning External Force Sensing for Commodity Robot Arms Improves Policy Learning