mlsanityData to 7 Oct 2026

AdaptNet: Policy Adaptation for Physics-Based Character Control

arXiv 2310.00239 ↗First posted 30 Sep 2023cs.GR cs.AI cs.LGcited by 15

Not on the map. Why?

Abstract (arXiv)

Motivated by humans' ability to adapt skills in the learning of new ones, this paper presents AdaptNet, an approach for modifying the latent space of existing policies to allow new behaviors to be quickly learned from like tasks in comparison to learning from scratch. Building on top of a given reinforcement learning controller, AdaptNet uses a two-tier hierarchy that augments the original state embedding to support modest changes in a behavior and further modifies the policy network layers to make more substantive changes. The technique is shown to be effective for adapting existing physics-based controllers to a wide range of new styles for locomotion, new task targets, changes in character morphology and extensive changes in environment. Furthermore, it exhibits significant increase in learning efficiency, as indicated by greatly reduced training times when compared to training from scratch or using other approaches that modify existing policies. Code is available at https://motion-lab.github.io/AdaptNet.

Key facts. Cited by 15 papers in the map (data through 7 Oct 2026). Builds on 61 of its 83 references that are in the map. First posted on arXiv 30 Sep 2023.

Builds on

61 of 83 references.

References · 61

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2017Proximal Policy Optimization Algorithmscited by 14,441
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  4. 2023Adding Conditional Control to Text-to-Image Diffusion Modelscited by 5,466
  5. 2017Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networkscited by 4,861
  6. 2019Analyzing and Improving the Image Quality of StyleGANcited by 3,580
  7. 2013Playing Atari with Deep Reinforcement Learningcited by 3,211
  8. 2015Domain-Adversarial Training of Neural Networkscited by 3,023
  9. 2014Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modelingcited by 2,341
  10. 2017Improved Training of Wasserstein GANscited by 2,223
  11. 2015Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networkscited by 2,048
  12. 2016Generative Adversarial Imitation Learningcited by 1,576
38 more
  1. 2023T2I-Adapter: Learning Adapters to Dig out More Controllable Ability for Text-to-Image Diffusion Modelscited by 1,339
  2. 2016Progressive Neural Networkscited by 1,332
  3. 2017Adversarial Discriminative Domain Adaptationcited by 1,309
  4. 2018On First-Order Meta-Learning Algorithmscited by 1,003
  5. 2021Isaac Gym: High Performance GPU-Based Physics Simulation For Robot Learningcited by 882
  6. 2019Interpreting the Latent Space of GANs for Semantic Face Editingcited by 691
  7. 2019Image2StyleGAN: How to Embed Images Into the StyleGAN Latent Space?cited by 675
  8. 2017Reinforcement Learning with Deep Energy-Based Policiescited by 659
  9. 2020Intrinsic Dimensionality Explains the Effectiveness of Language Model Fine-Tuningcited by 658
  10. 2016Learning to learn by gradient descent by gradient descentcited by 616
  11. 2017Sim-to-Real Transfer of Robotic Control with Dynamics Randomizationcited by 616
  12. 2016Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modelingcited by 558
  13. 2016RL2: Fast Reinforcement Learning via Slow Reinforcement Learningcited by 490
  14. 2015Policy Distillationcited by 414
  15. 2018Measuring the Intrinsic Dimension of Objective Landscapescited by 364
  16. 2021The Intrinsic Dimension of Images and Its Impact on Learningcited by 307
  17. 2019On the "steerability" of generative adversarial networkscited by 247
  18. 2017Geometric GANcited by 232
  19. 2021Robust Motion In-betweeningcited by 223
  20. 2021Character Controllers Using Motion VAEscited by 182
  21. 2015Actor-Mimic: Deep Multitask and Transfer Reinforcement Learningcited by 176
  22. 2017Emergence of Locomotion Behaviours in Rich Environmentscited by 159
  23. 2017Optimizing the Latent Space of Generative Networkscited by 157
  24. 2018Meta-Gradient Reinforcement Learningcited by 156
  25. 2018DeepMimic: Example-Guided Deep Reinforcement Learning of Physics-Based Character Skillscited by 153
  26. 2018Residual Policy Learningcited by 146
  27. 2016EPOpt: Learning Robust Neural Network Policies Using Model Ensemblescited by 136
  28. 2018Meta-Reinforcement Learning of Structured Exploration Strategiescited by 128
  29. 2016Learning Modular Neural Network Policies for Multi-Task and Multi-Robot Transfercited by 124
  30. 2017BEGAN: Boundary Equilibrium Generative Adversarial Networkscited by 121
  31. 2019MCP: Learning Composable Hierarchical Control with Multiplicative Compositional Policiescited by 105
  32. 2020Unpaired Motion Style Transfer from Video to Animationcited by 94
  33. 2023CALM: Conditional Adversarial Latent Models for Directable Virtual Characterscited by 91
  34. 2016Sim-to-Real Robot Learning from Pixels with Progressive Netscited by 89
  35. 2018Evolved Policy Gradientscited by 86
  36. 2015Memory-based control with recurrent neural networkscited by 77
  37. 2017Learning human behaviors from motion capture by adversarial imitationcited by 75
  38. 2018Learning Symmetric and Low-energy Locomotioncited by 53

50 most cited of 61 All 61 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 5; 2025: 5; 2026: 5.

15 papers.

Sort:

By area:

  1. 2025TokenHSI: Unified Synthesis of Physical Human-Scene Interactions through Task Tokenizationcited by 46
  2. 2024SMooDi: Stylized Motion Diffusion Modelcited by 37
  3. 2024PACER+: On-Demand Pedestrian Animation Controller in Driving Scenarioscited by 24
  4. 2024Lodge++: High-quality and Long Dance Generation with Vivid Choreography Patternscited by 21
  5. 2025Learning to Ball: Composing Policies for Long-Horizon Basketball Movescited by 6
  6. 2024From Words to Worlds: Transforming One-line Prompt into Immersive Multi-modal Digital Stories with Communicative LLM Agentcited by 4
  7. 2025CBIL: Collective Behavior Imitation Learning for Fish from Real Videoscited by 4
  8. 2024ChatDyn: Language-Driven Multi-Actor Dynamics Generation in Street Scenescited by 3
  9. 2025Versatile Physics-based Character Control with Hybrid Latent Representationcited by 3
  10. 2025Environment-aware Motion Matchingcited by 1
  11. 2026MUSIC: Learning Muscle-Driven Dexterous Hand Controlcited by 1
  12. 2026IAM: Identity-Aware Human Motion and Shape Joint Generationcited by 1
3 more
  1. 2026MaskAdapt: Learning Flexible Motion Adaptation via Mask-Invariant Prior for Physics-Based Characterscited by 0
  2. 2026NaP-Control: Navigating Diffusion Prior for Versatile and Fast Character Controlcited by 0
  3. 2026Policy-as-Data: Learning Generalizable HOI Diffusion Models from Simulated Physicscited by 0