Meta Inverse Reinforcement Learning via Maximum Reward Sharing for Human Motion Analysis
arXiv:1710.03592
Abstract
This work handles the inverse reinforcement learning (IRL) problem where only a small number of demonstrations are available from a demonstrator for each high-dimensional task, insufficient to estimate an accurate reward function. Observing that each demonstrator has an inherent reward for each state and the task-specific behaviors mainly depend on a small number of key states, we propose a meta IRL algorithm that first models the reward function for each task as a distribution conditioned on a baseline reward function shared by all tasks and dependent only on the demonstrator, and then finds the most likely reward function in the distribution that explains the task-specific behaviors. We test the method in a simulated environment on path planning tasks with limited demonstrations, and show that the accuracy of the learned reward function is significantly improved. We also apply the method to analyze the motion of a patient under rehabilitation.
arXiv commentarXiv admin note: text overlap with arXiv:1707.09394
References in corpus (7)
- Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
- RL: Fast Reinforcement Learning via Slow Reinforcement Learning
- Maximum Entropy Deep Inverse Reinforcement Learning
- One-Shot Imitation Learning
- Actor-Mimic: Deep Multitask and Transfer Reinforcement Learning
- Bayesian multitask inverse reinforcement learning
- Inverse Reinforcement Learning in Large State Spaces via Function Approximation