Universal Successor Representations for Transfer Reinforcement Learning
This work addresses transfer learning in reinforcement learning for scenarios with fixed dynamics and varying goals, offering an incremental improvement over existing methods like general value functions.
The paper tackled the problem of transfer reinforcement learning where tasks share dynamics but have different goals, proposing universal successor representations (USR) and an approximator (USRA) to represent transferable knowledge. The result showed that agents initialized with USRA achieved goals faster than random initialization, with concrete improvements in training speed.
The objective of transfer reinforcement learning is to generalize from a set of previous tasks to unseen new tasks. In this work, we focus on the transfer scenario where the dynamics among tasks are the same, but their goals differ. Although general value function (Sutton et al., 2011) has been shown to be useful for knowledge transfer, learning a universal value function can be challenging in practice. To attack this, we propose (1) to use universal successor representations (USR) to represent the transferable knowledge and (2) a USR approximator (USRA) that can be trained by interacting with the environment. Our experiments show that USR can be effectively applied to new tasks, and the agent initialized by the trained USRA can achieve the goal considerably faster than random initialization.