Classifying Options for Deep Reinforcement Learning
This work addresses sample efficiency in reinforcement learning for tasks with mixed transfer effects, but it is incremental as it builds on existing methods like DQNs and the options framework.
The paper tackles the problem of hierarchical reinforcement learning by combining the options framework with deep Q-networks using option heads and a supervisory network, resulting in lower sample complexity for subtasks with negative transfer without degrading performance for positive transfer.
In this paper we combine one method for hierarchical reinforcement learning - the options framework - with deep Q-networks (DQNs) through the use of different "option heads" on the policy network, and a supervisory network for choosing between the different options. We utilise our setup to investigate the effects of architectural constraints in subtasks with positive and negative transfer, across a range of network capacities. We empirically show that our augmented DQN has lower sample complexity when simultaneously learning subtasks with negative transfer, without degrading performance when learning subtasks with positive transfer.