mlsanityData to 7 Oct 2026

LOTUS: Continual Imitation Learning for Robot Manipulation Through Unsupervised Skill Discovery

arXiv 2311.02058 ↗First posted 3 Nov 2023cs.RO cs.CV cs.LGcited by 63

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

Abstract (arXiv)

We introduce LOTUS, a continual imitation learning algorithm that empowers a physical robot to continuously and efficiently learn to solve new manipulation tasks throughout its lifespan. The core idea behind LOTUS is constructing an ever-growing skill library from a sequence of new tasks with a small number of human demonstrations. LOTUS starts with a continual skill discovery process using an open-vocabulary vision model, which extracts skills as recurring patterns presented in unsegmented demonstrations. Continual skill discovery updates existing skills to avoid catastrophic forgetting of previous tasks and adds new skills to solve novel tasks. LOTUS trains a meta-controller that flexibly composes various skills to tackle vision-based manipulation tasks in the lifelong learning process. Our comprehensive experiments show that LOTUS outperforms state-of-the-art baselines by over 11% in success rate, showing its superior knowledge transfer ability compared to prior methods. More results and videos can be found on the project website: https://ut-austin-rpl.github.io/Lotus/.

Key facts. Cited by 63 papers in the map (data through 7 Oct 2026). Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data. Builds on 47 of its 63 references that are in the map. First posted on arXiv 3 Nov 2023.

Builds on

47 of 63 references.

Closest ancestors · 10

Same direction, newest first

  1. 2023Learning Generalizable Manipulation Policies with Object-Centric 3D Representations22 Oct
  2. 2023Multi-Stage Cable Routing through Hierarchical Imitation Learning18 Jul
  3. 2023MimicPlay: Long-Horizon Imitation Learning by Watching Human Play24 Feb
  4. 2022VIOLA: Imitation Learning for Vision-Based Manipulation with Object Proposal Priors20 Oct
  5. 2021Bottom-Up Skill Discovery from Unsegmented Demonstrations for Long-Horizon Robot Manipulation28 Sep
  6. 2021CRIL: Continual Robot Imitation Learning via Generative and Prediction Model17 Jun
  7. 2020Learning Multi-Arm Manipulation Through Collaborative Teleoperation12 Dec
  8. 2020Learning to Generalize Across Long-Horizon Tasks from Human Demonstrations13 Mar
  9. 2019Relay Policy Learning: Solving Long-Horizon Tasks via Imitation and Reinforcement Learning25 Oct
  10. 2018One-Shot Hierarchical Imitation Learning of Compound Visuomotor Tasks25 Oct

Other references · 37

  1. 2017Attention Is All You Needcited by 59,932
  2. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  3. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  4. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  5. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  6. 2016Overcoming catastrophic forgetting in neural networkscited by 4,718
  7. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,730
  8. 2023Voyager: An Open-Ended Embodied Agent with Large Language Modelscited by 2,081
  9. 2017Gradient Episodic Memory for Continual Learningcited by 1,871
  10. 2023LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learningcited by 1,338
  11. 2023A Comprehensive Survey of Continual Learning: Theory, Method and Applicationcited by 860
  12. 2018Diversity is All You Need: Learning Skills without a Reward Functioncited by 734
25 more
  1. 2021What Matters in Learning from Offline Human Demonstrations for Robot Manipulationcited by 699
  2. 2022R3M: A Universal Visual Representation for Robot Manipulationcited by 678
  3. 2017FeUdal Networks for Hierarchical Reinforcement Learningcited by 328
  4. 2019Learning Latent Plans from Playcited by 310
  5. 2019Dynamics-Aware Unsupervised Discovery of Skillscited by 308
  6. 2016Variational Intrinsic Controlcited by 262
  7. 2018Selective Experience Replay for Lifelong Learningcited by 239
  8. 2018Don't forget, there is more than forgetting: new metrics for Continual Learningcited by 113
  9. 2016A Deep Hierarchical Approach to Lifelong Learning in Minecraftcited by 90
  10. 2018Hierarchical Imitation and Reinforcement Learningcited by 88
  11. 2019IRIS: Implicit Reinforcement without Interaction at Scale for Learning Control from Offline Robot Manipulation Datacited by 85
  12. 2023MUTEX: Learning Unified Policies from Multimodal Task Specificationscited by 70
  13. 2018TACO: Learning Task Decomposition via Temporal Alignment for Controlcited by 61
  14. 2017Multi-Level Discovery of Deep Optionscited by 50
  15. 2020Deep Reinforcement Learning amidst Lifelong Non-Stationaritycited by 42
  16. 2022Modular Lifelong Reinforcement Learning via Neural Compositioncited by 39
  17. 2021Lifelong Robotic Reinforcement Learning by Retaining Experiencescited by 34
  18. 2022Lifelong Reinforcement Learning with Modulating Maskscited by 20
  19. 2022Fast Lifelong Adaptive Inverse Reinforcement Learning from Demonstrationscited by 14
  20. 2023PolyTask: Learning Unified Policies through Behavior Distillationcited by 11
  21. 2022One After Another: Learning Incremental Skills for a Changing Worldcited by 10
  22. 2022Continual Vision-based Reinforcement Learning with Group Symmetriescited by 10
  23. 2022Forgetting and Imbalance in Robot Lifelong Learning with Off-policy Datacited by 5
  24. 2020Lifelong Learning using Eigentasks: Task Separation, Skill Acquisition, and Selective Transfercited by 4
  25. 2023Evaluating Continual Learning on a Home Robotcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 1; 2024: 11; 2025: 24; 2026: 27.

63 papers.

Sort:

By area:

  1. 2024Sparse Diffusion Policy: A Sparse, Reusable, and Flexible Policy for Robot Learningcited by 56
  2. 2024Deep Generative Models in Robotics: A Survey on Learning from Multimodal Demonstrationscited by 35
  3. 2024Practice Makes Perfect: Planning to Learn Skill Parameter Policiescited by 22
  4. 2024Incremental Learning of Retrievable Skills For Efficient Continual Task Adaptationcited by 19
  5. 2025Robust Finetuning of Vision-Language-Action Robot Policies via Parameter Mergingcited by 19
  6. 2025Dexterous Manipulation through Imitation Learning: A Surveycited by 18
  7. 2026Playful Agentic Robot Learningcited by 18
  8. 2025Learning Compositional Behaviors from Demonstration and Languagecited by 15
  9. 2026Pretrained Vision-Language-Action Models are Surprisingly Resistant to Forgetting in Continual Learningcited by 15
  10. 2026AtomicVLA: Unlocking the Potential of Atomic Skill Learning in Robotscited by 15
  11. 2026CLARE: Continual Learning for Vision-Language-Action Models via Autonomous Adapter Routing and Expansioncited by 14
  12. 2024EXTRACT: Efficient Policy Learning by Extracting Transferable Robot Skills from Offline Datacited by 13
38 more
  1. 2025Think Small, Act Big: Primitive Prompt Learning for Lifelong Robot Manipulationcited by 13
  2. 2025A Survey on Imitation Learning for Contact-Rich Tasks in Roboticscited by 12
  3. 2026Towards Long-Lived Robots: Continual Learning VLA Models via Reinforcement Fine-Tuningcited by 11
  4. 2024Never-Ending Behavior-Cloning Agent for Robotic Manipulationcited by 10
  5. 2024M2Distill: Multi-Modal Distillation for Lifelong Imitation Learningcited by 10
  6. 2025X-IL: Exploring the Design Space of Imitation Learning Policiescited by 9
  7. 2025SPECI: Skill Prompts based Hierarchical Continual Imitation Learning for Robot Manipulationcited by 9
  8. 2025Continual Learning for Generative AI: From LLMs to MLLMs and Beyondcited by 9
  9. 2025Dynamic Mixture of Progressive Parameter-Efficient Expert Library for Lifelong Robot Learningcited by 8
  10. 2024DABI: Evaluation of Data Augmentation Methods Using Downsampling in Bilateral Control-Based Imitation Learning with Imagescited by 7
  11. 2026GuidedVLA: Specifying Task-Relevant Factors via Plug-and-Play Action Attention Specializationcited by 7
  12. 2023Scalable and Efficient Continual Learning from Demonstration via a Hypernetwork-generated Stable Dynamics Modelcited by 6
  13. 2024Solving Continual Offline RL through Selective Weights Activation on Aligned Spacescited by 6
  14. 2025Continual Adaptation for Autonomous Driving with the Mixture of Progressive Experts Networkcited by 4
  15. 2025ADPro: a Test-time Adaptive Diffusion Policy via Manifold-constrained Denoising and Task-aware Initialization for Robotic Manipulationcited by 4
  16. 2026DreamProver: Evolving Transferable Lemma Libraries via a Wake-Sleep Theorem-Proving Agentcited by 4
  17. 2024Composing Diffusion Policies for Few-shot Learning of Movement Trajectoriescited by 3
  18. 2025Few-Shot Vision-Language Action-Incremental Policy Learningcited by 3
  19. 2025Scalable Strategies for Continual Learning with Replaycited by 3
  20. 2025Robotic Manipulation via Imitation Learning: Taxonomy, Evolution, Benchmark, and Challengescited by 3
  21. 2025LodeStar: Long-horizon Dexterity via Synthetic Data Augmentation from Human Demonstrationscited by 3
  22. 2025VER: Vision Expert Transformer for Robot Learning via Foundation Distillation and Dynamic Routingcited by 3
  23. 2026Lifelong Embodied Navigation Learningcited by 3
  24. 2025Sample-Efficient Behavior Cloning Using General Domain Knowledgecited by 2
  25. 2025T2S: Tokenized Skill Scaling for Lifelong Imitation Learningcited by 2
  26. 2025Policy Compatible Skill Incremental Learning via Lazy Learning Interfacecited by 2
  27. 2026WeaveLA: Event Driven Cross-Subtask Latent Memory Weaving for Repetitive Robot Manipulationcited by 2
  28. 2026World Action Models Enable Continual Imitation Learning with Recurrent Generative Replayscited by 2
  29. 2024Task-agnostic Lifelong Robot Learning with Retrieval-based Weighted Local Adaptationcited by 1
  30. 2025Casper: Inferring Diverse Intents for Assistive Teleoperation with Vision Language Modelscited by 1
  31. 2025Learning Parameterized Skills from Demonstrationscited by 1
  32. 2026SPREAD: Subspace Representation Distillation for Lifelong Imitation Learningcited by 1
  33. 2026Lifelong Imitation Learning with Multimodal Latent Replay and Incremental Adjustmentcited by 1
  34. 2026All-day Multi-scenes Lifelong Vision-and-Language Navigation with Tucker Adaptationcited by 1
  35. 2026PHASER: Phase-Aware and Semantic Experience Replay for Vision-Language-Action Modelscited by 1
  36. 2026Learning New Tasks via Reusable Skills: Skill-Compositional Experts for Embodied Continual Learningcited by 1
  37. 2026LiMoDE: Rethinking Lifelong Robot Manipulation from a Mixture-of-Dynamic-Experts Perspectivecited by 1
  38. 2025Coupled Distributional Random Expert Distillation for World Model Online Imitation Learningcited by 0

50 most cited of 63 All 63 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2022Learning and Retrieval from Prior Data for Skill-based Imitation Learning
  2. 2026Beyond the Majority: Long-tail Imitation Learning for Robotic Manipulation
  3. 2025Learning Semantic Atomic Skills for Multi-Task Robotic Manipulation
  4. 2020Scalable Multi-Task Imitation Learning with Autonomous Improvement
  5. 2026Human-to-Robot Interaction: Learning from Video Demonstration for Robot Imitation
  6. 2026Offline Discovery of Interpretable Skills from Multi-Task Trajectories
  7. 2021Towards More Generalizable One-shot Visual Imitation Learning
  8. 2023Teach a Robot to FISH: Versatile Imitation from One Minute of Demonstrations