Ashwin Balakrishna

Latest

• Dynamics-Aware Comparison of Learned Reward Functions
• Safety Augmented Value Estimation from Demonstrations (SAVED): Safe Deep Model-Based RL for Sparse Cost Robotic Tasks