mlsanityData to 14 Oct 2026

Data Scaling Laws in Imitation Learning for Robotic Manipulation

arXiv 2410.18647 ↗First posted 24 Oct 2024cs.ROcited by 145

On the map: Offline RL and imitation learning from data → Robot manipulation imitation learning

Abstract (arXiv)

Data scaling has revolutionized fields like natural language processing and computer vision, providing models with remarkable generalization capabilities. In this paper, we investigate whether similar data scaling laws exist in robotics, particularly in robotic manipulation, and whether appropriate data scaling can yield single-task robot policies that can be deployed zero-shot for any object within the same category in any environment. To this end, we conduct a comprehensive empirical study on data scaling in imitation learning. By collecting data across numerous environments and objects, we study how a policy's generalization performance changes with the number of training environments, objects, and demonstrations. Throughout our research, we collect over 40,000 demonstrations and execute more than 15,000 real-world robot rollouts under a rigorous evaluation protocol. Our findings reveal several intriguing results: the generalization performance of the policy follows a roughly power-law relationship with the number of environments and objects. The diversity of environments and objects is far more important than the absolute number of demonstrations; once the number of demonstrations per environment or object reaches a certain threshold, additional demonstrations have minimal effect. Based on these insights, we propose an efficient data collection strategy. With four data collectors working for one afternoon, we collect sufficient data to enable the policies for two tasks to achieve approximately 90% success rates in novel environments with unseen objects.

Key facts. Cited by 145 papers in the map (data through 14 Oct 2026). Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data. Builds on 70 of its 80 references that are in the map. First posted on arXiv 24 Oct 2024.

Builds on

70 of 80 references.

Closest ancestors · 17

Same direction, newest first

  1. 2024ALOHA Unleashed: A Simple Recipe for Robot Dexterity17 Oct
  2. 2024Open-TeleVision: Teleoperation with Immersive Active Visual Feedback1 Jul
  3. 2024Mobile ALOHA: Learning Bimanual Mobile Manipulation with Low-Cost Whole-Body Teleoperation4 Jan
  4. 2023Any-point Trajectory Modeling for Policy Learning28 Dec
  5. 2023Learning Generalizable Manipulation Policies with Object-Centric 3D Representations22 Oct
  6. 2023GELLO: A General, Low-Cost, and Intuitive Teleoperation Framework for Robot Manipulators22 Sep
  7. 2023BridgeData V2: A Dataset for Robot Learning at Scale24 Aug
  8. 2023Decomposing the Generalization Gap in Imitation Learning for Visual Robotic Manipulation7 Jul
  9. 2023Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware23 Apr
  10. 2023MimicPlay: Long-Horizon Imitation Learning by Watching Human Play24 Feb
  11. 2022VIOLA: Imitation Learning for Vision-Based Manipulation with Object Proposal Priors20 Oct
  12. 2022VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training30 Sep
5 more
  1. 2022BC-Z: Zero-Shot Task Generalization with Robotic Imitation Learning4 Feb
  2. 2021The Surprising Effectiveness of Representation Learning for Visual Imitation2 Dec
  3. 2021DexMV: Imitation Learning for Dexterous Manipulation from Human Videos12 Aug
  4. 2020Visual Imitation Made Easy11 Aug
  5. 2018RoboTurk: A Crowdsourcing Platform for Robotic Skill Learning through Imitation7 Nov

Other references · 53

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  3. 2020Language Models are Few-Shot Learnerscited by 32,419
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  5. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  6. 2023GPT-4 Technical Reportcited by 18,396
  7. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  8. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  9. 2020Denoising Diffusion Implicit Modelscited by 9,781
  10. 2023Segment Anythingcited by 8,234
  11. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  12. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,403
38 more
  1. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  2. 2020Scaling Laws for Neural Language Modelscited by 6,586
  3. 2022Scalable Diffusion Models with Transformerscited by 6,054
  4. 2021On the Opportunities and Risks of Foundation Modelscited by 4,033
  5. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 3,080
  6. 2024OpenVLA: An Open-Source Vision-Language-Action Modelcited by 2,887
  7. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,838
  8. 2022Do As I Can, Not As I Say: Grounding Language in Robotic Affordancescited by 2,348
  9. 2022RT-1: Robotics Transformer for Real-World Control at Scalecited by 2,061
  10. 2024Octo: An Open-Source Generalist Robot Policycited by 1,484
  11. 2024Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Predictioncited by 1,023
  12. 2021Scaling Vision Transformerscited by 999
  13. 2023Open X-Embodiment: Robotic Learning Datasets and RT-X Modelscited by 862
  14. 2024DROID: A Large-Scale In-The-Wild Robot Manipulation Datasetcited by 852
  15. 2018QT-Opt: Scalable Deep Reinforcement Learning for Vision-Based Robotic Manipulationcited by 743
  16. 2020Scaling Laws for Autoregressive Generative Modelingcited by 522
  17. 2020robosuite: A Modular Simulation Framework and Benchmark for Robot Learningcited by 511
  18. 2024Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robotscited by 394
  19. 2017Deep Imitation Learning for Complex Manipulation Tasks from Virtual Reality Teleoperationcited by 289
  20. 2019RoboNet: Large-Scale Multi-Robot Learningcited by 270
  21. 2017Dex-Net 2.0: Deep Learning to Plan Robust Grasps with Synthetic Point Clouds and Analytic Grasp Metricscited by 264
  22. 2021Bridge Data: Boosting Generalization of Robotic Skills with Cross-Domain Datasetscited by 250
  23. 2022AnyGrasp: Robust and Efficient Grasp Perception in Spatial and Temporal Domainscited by 208
  24. 2021ManiSkill: Generalizable Manipulation Skill Benchmark with Large-Scale Demonstrationscited by 197
  25. 2023RH20T: A Comprehensive Robotic Dataset for Learning Diverse Skills in One-Shotcited by 183
  26. 2023RoboAgent: Generalization and Efficiency in Robot Manipulation via Semantic Augmentations and Action Chunkingcited by 169
  27. 2022Human-to-Robot Imitation in the Wildcited by 161
  28. 2023Look Before You Leap: Unveiling the Power of GPT-4V in Robotic Vision-Language Planningcited by 161
  29. 2024THE COLOSSEUM: A Benchmark for Evaluating Generalization for Robotic Manipulationcited by 133
  30. 2020Self-Supervised Policy Adaptation during Deploymentcited by 123
  31. 2021Learning Generalizable Robotic Reward Functions from "In-The-Wild" Human Videoscited by 119
  32. 2019Grasping in the Wild:Learning 6DoF Closed-Loop Grasping from Low-Cost Demonstrationscited by 107
  33. 2023On Bringing Robots Homecited by 101
  34. 2018Multiple Interactions Made Easy (MIME): Large Scale Demonstrations Data for Imitationcited by 76
  35. 2024OPEN TEACH: A Versatile Teleoperation System for Robotic Manipulationcited by 71
  36. 2024Vision-based Manipulation from Single Human Video with Open-World Object Graphscited by 50
  37. 2024Robot Utility Models: General Policies for Zero-Shot Deployment in New Environmentscited by 44
  38. 2023Scaling laws for single-agent reinforcement learningcited by 41

50 most cited of 53 All 53 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 10; 2025: 73; 2026: 62.

145 papers.

Sort:

By area:

  1. 2025π0.5: a Vision-Language-Action Model with Open-World Generalizationcited by 1,761
  2. 2025CoT-VLA: Visual Chain-of-Thought Reasoning for Vision-Language-Action Modelscited by 451
  3. 2025AgiBot World Colosseo: A Large-scale Manipulation Platform for Scalable and Intelligent Embodied Systemscited by 405
  4. 2026Causal World Modeling for Robot Controlcited by 321
  5. 2025X-VLA: Soft-Prompted Transformer as Scalable Cross-Embodiment Vision-Language-Action Modelcited by 303
  6. 2025DreamVLA: A Vision-Language-Action Model Dreamed with Comprehensive World Knowledgecited by 213
  7. 2025Unified Video Action Modelcited by 205
  8. 2025DexVLA: Vision-Language Model with Plug-In Diffusion Expert for General Robot Controlcited by 195
  9. 2025VLA-RL: Towards Masterful and General Robotic Manipulation with Scalable Reinforcement Learningcited by 176
  10. 2024VLABench: A Large-Scale Benchmark for Language-Conditioned Robotics Manipulation with Long-Horizon Reasoning Taskscited by 175
  11. 2025DriveVLA-W0: World Models Amplify Data Scaling Law in Autonomous Drivingcited by 134
  12. 2026VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Modelcited by 118
38 more
  1. 2025ChatVLA: Unified Multimodal Understanding and Robot Control with Vision-Language-Action Modelcited by 113
  2. 2025GR-3 Technical Reportcited by 111
  3. 2025OneTwoVLA: A Unified Vision-Language-Action Model with Adaptive Reasoningcited by 107
  4. 2025ReinFlow: Fine-tuning Flow Matching Policy with Online Reinforcement Learningcited by 97
  5. 2025A Vision-Language-Action-Critic Model for Robotic Real-World Reinforcement Learningcited by 71
  6. 2025WMPO: World Model-based Policy Optimization for Vision-Language-Action Modelscited by 70
  7. 2024Diffusion-VLA: Generalizable and Interpretable Robot Foundation Model via Self-Generated Reasoningcited by 57
  8. 2025FLOWER: Democratizing Generalist Robot Policies with Efficient Vision-Language-Action Flow Policiescited by 57
  9. 2026EgoVerse: An Egocentric Human Dataset for Robot Learning from Around the Worldcited by 57
  10. 2025Training Strategies for Efficient Embodied Reasoningcited by 55
  11. 2025Sim-to-Real Reinforcement Learning for Vision-Based Dexterous Manipulation on Humanoidscited by 48
  12. 2025DynaGuide: Steering Diffusion Polices with Active Dynamic Guidancecited by 44
  13. 2025Failure Prediction at Runtime for Generative Robot Policiescited by 42
  14. 2025RaC: Robot Learning for Long-Horizon Tasks by Scaling Recovery and Correctioncited by 39
  15. 2025DexWild: Dexterous Human Interactions for In-the-Wild Robot Policiescited by 38
  16. 2025Embodied-R1: Reinforced Embodied Reasoning for General Robotic Manipulationcited by 38
  17. 2025Much Ado About Noising: Dispelling the Myths of Generative Robotic Controlcited by 36
  18. 2024G3Flow: Generative 3D Semantic Flow for Pose-aware and Generalizable Object Manipulationcited by 35
  19. 2025From Seeing to Doing: Bridging Reasoning and Decision for Robotic Manipulationcited by 35
  20. 2025FP3: A 3D Foundation Policy for Robotic Manipulationcited by 32
  21. 2026World-Gymnast: Training Robots with Reinforcement Learning in a World Modelcited by 27
  22. 2025Robobench: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models as Embodied Braincited by 25
  23. 2024RoboMatrix: A Skill-centric Hierarchical Framework for Scalable Robot Task Planning and Execution in Open-Worldcited by 22
  24. 2025PIN-WM: Learning Physics-INformed World Models for Non-Prehensile Manipulationcited by 22
  25. 2025SKIL: Semantic Keypoint Imitation Learning for Generalizable Data-efficient Manipulationcited by 21
  26. 2026PALM: Progress-Aware Policy Learning via Affordance Reasoning for Long-Horizon Robotic Manipulationcited by 21
  27. 2025Is Diversity All You Need for Scalable Robotic Manipulation?cited by 20
  28. 2025AirExo-2: Scaling up Generalizable Robotic Imitation Learning with Low-Cost Exoskeletonscited by 18
  29. 2026LAP: Language-Action Pre-Training Enables Zero-shot Cross-Embodiment Transfercited by 18
  30. 2025ORV: 4D Occupancy-centric Robot Video Generationcited by 17
  31. 2025Parallels Between VLA Model Post-Training and Human Motor Learning: Progress, Challenges, and Trendscited by 17
  32. 2026CLARE: Continual Learning for Vision-Language-Action Models via Autonomous Adapter Routing and Expansioncited by 17
  33. 2025A Taxonomy for Evaluating Generalist Robot Manipulation Policiescited by 16
  34. 2025A Survey of Robotic Navigation and Manipulation with Physics Simulators in the Era of Embodied AIcited by 16
  35. 2025Towards Embodiment Scaling Laws in Robot Locomotioncited by 16
  36. 2025Do You Need Proprioceptive States in Visuomotor Policies?cited by 16
  37. 2025Normalizing Flows are Capable Models for Continuous Controlcited by 14
  38. 2025Touch begins where vision ends: Generalizable policies for contact-rich manipulationcited by 14

50 most cited of 145 All 145 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026The Curse of Precision: A Data Scaling Law for High-Precision Robotic Manipulation
  2. 2025What Matters in Learning from Large-Scale Datasets for Robot Manipulation
  3. 2026Training and Evaluating Diffusion Policies with Long Context Lengths
  4. 2025Guiding Data Collection via Factored Scaling Curves
  5. 2026Beyond the Majority: Long-tail Imitation Learning for Robotic Manipulation
  6. 2025MOVE: A Simple Motion-Based Data Collection Paradigm for Spatial Generalization in Robotic Manipulation
  7. 2025Robot Data Curation with Mutual Information Estimators
  8. 2026A Systematic Study of Data Modalities and Strategies for Co-training Large Behavior Models for Robot Manipulation