Jan-Willem van de Meent

Latest

• On-Robot Reinforcement Learning with Goal-Contrastive Rewards
• One-shot Imitation Learning via Interaction Warping
• Learning Symmetric Embeddings for Equivariant World Models
• Binding Actions to Objects in World Models
• Binding Actions to Objects in World Models
• Action Priors for Large Action Spaces in Robotics
• Learning Discrete State Abstractions With Deep Variational Inference