Leveraging Human Guidance for Deep Reinforcement Learning Tasks
arXiv:1909.09906
Abstract
Reinforcement learning agents can learn to solve sequential decision tasks by interacting with the environment. Human knowledge of how to solve these tasks can be incorporated using imitation learning, where the agent learns to imitate human demonstrated decisions. However, human guidance is not limited to the demonstrations. Other types of guidance could be more suitable for certain tasks and require less human effort. This survey provides a high-level overview of five recent learning frameworks that primarily rely on human guidance other than conventional, step-by-step action demonstrations. We review the motivation, assumption, and implementation of each framework. We then discuss possible future research directions.
Proceedings of the 28th International Joint Conference on Artificial Intelligence (IJCAI 2019)
References in corpus (12)
- Deep reinforcement learning from human preferences
- Generative Adversarial Imitation Learning
- Deep Reinforcement Learning that Matters
- Modular Multitask Reinforcement Learning with Policy Sketches
- Learning human behaviors from motion capture by adversarial imitation
- Third-Person Imitation Learning
- Learning Invariant Feature Spaces to Transfer Skills with Reinforcement Learning
- Generative Adversarial Imitation from Observation
- Imitating Latent Policies from Observation
- DQN-TAMER: Human-in-the-Loop Reinforcement Learning with Intractable Feedback
- Atari-HEAD: Atari Human Eye-Tracking and Demonstration Dataset
- Learning Actionable Representations from Visual Observations