Learning human behaviors from motion capture by adversarial imitation

Abstract

Rapid progress in deep reinforcement learning has made it increasingly feasible to train controllers for high-dimensional humanoid bodies. However, methods that use pure reinforcement learning with simple reward functions tend to produce non-humanlike and overly stereotyped movement behaviors. In this work, we extend generative adversarial imitation learning to enable training of generic neural network policies to produce humanlike movement patterns from limited demonstrations consisting only of partially observed state features, without access to actions, even when the demonstrations come from a body with different and unknown physical parameters. We leverage this approach to build sub-skill policies from motion capture data and show that they can be reused to solve tasks when controlled by a higher level controller.

Cited by in corpus (51)

• DeepMimic: Example-Guided Deep Reinforcement Learning of Physics-Based Character Skills
• AMP: Adversarial Motion Priors for Stylized Physics-Based Character Control
• A Survey of Deep RL and IL for Autonomous Driving Policy Learning
• dm_control: Software and Tasks for Continuous Control
• Variational Discriminator Bottleneck: Improving Imitation Learning, Inverse RL, and GANs by Constraining Information Flow
• Reinforcement and Imitation Learning for Diverse Visuomotor Skills
• Generative Adversarial Imitation from Observation
• Transflower: probabilistic autoregressive dance generation with multimodal attention
• Hierarchical visuomotor control of humanoids
• Robust Imitation of Diverse Behaviors
• Making Efficient Use of Demonstrations to Solve Hard Exploration Problems
• Scaling data-driven robotics with reward sketching and batch reinforcement learning
• Imitating Interactive Intelligence
• Residual Force Control for Agile Human Behavior Imitation and Extended Motion Synthesis
• PADL: Language-Directed Physics-Based Character Control
• Composite Motion Learning with Task Control
• Progressive Reinforcement Learning with Distillation for Multi-Skilled Motion Control
• Feedback Control For Cassie With Deep Reinforcement Learning
• Toward the Fundamental Limits of Imitation Learning
• Learning to Generate Pointing Gestures in Situated Embodied Conversational Agents
• Offline Learning from Demonstrations and Unlabeled Experience
• UniCon: Universal Neural Controller For Physics-based Character Motion
• Ego-Pose Estimation and Forecasting as Real-Time PD Control
• Demonstration-Guided Reinforcement Learning with Learned Skills
• Situated GAIL: Multitask imitation using task-conditioned adversarial inverse reinforcement learning
• Semi-supervised reward learning for offline reinforcement learning
• Learning to Sit: Synthesizing Human-Chair Interactions via Hierarchical Control
• Adversarial Imitation Learning from Incomplete Demonstrations
• Triple-GAIL: A Multi-Modal Imitation Learning Framework with Generative Adversarial Nets
• Generative Adversarial Imitation Learning with Neural Networks: Global Optimality and Convergence Rate
• Leveraging Human Guidance for Deep Reinforcement Learning Tasks
• Reinforced Imitation in Heterogeneous Action Space
• Domain-Adversarial and Conditional State Space Model for Imitation Learning
• SimPoE: Simulated Character Control for 3D Human Pose Estimation
• Probabilistic model predictive safety certification for learning-based control
• Evaluation metrics for behaviour modeling
• Provably Efficient Generative Adversarial Imitation Learning for Online and Offline Setting with Linear Function Approximation
• Hierarchical Skills for Efficient Exploration
• PFPN: Continuous Control of Physically Simulated Characters using Particle Filtering Policy Network
• On the Guaranteed Almost Equivalence between Imitation Learning from Observation and Demonstration
• NEARL: Non-Explicit Action Reinforcement Learning for Robotic Control
• HILONet: Hierarchical Imitation Learning from Non-Aligned Observations
• No Need for Interactions: Robust Model-Based Imitation Learning using Neural ODE
• Integration of Imitation Learning using GAIL and Reinforcement Learning using Task-achievement Rewards via Probabilistic Graphical Model
• Towards Learning to Imitate from a Single Video Demonstration
• A Bayesian Approach to Identifying Representational Errors
• Autonomous Functional Locomotion in a Tendon-Driven Limb via Limited Experience
• Injective State-Image Mapping facilitates Visual Adversarial Imitation Learning
• Relational Mimic for Visual Adversarial Imitation Learning
• GRP Model for Sensorimotor Learning
• Procedure Planning in Instructional Videos via Contextual Modeling and Model-based Policy Learning