Improving generalization in reinforcement learning through forked agents
This work addresses generalization challenges in reinforcement learning for AI systems, but it appears incremental as it builds on existing eco-system approaches with new initialization methods.
The paper tackled the problem of improving generalization in reinforcement learning across procedurally generated environments by proposing different initialization techniques for new agents in an eco-system approach, resulting in a study of their impact on adaptation speed and effectiveness.
An eco-system of agents each having their own policy with some, but limited, generalizability has proven to be a reliable approach to increase generalization across procedurally generated environments. In such an approach, new agents are regularly added to the eco-system when encountering a new environment that is outside of the scope of the eco-system. The speed of adaptation and general effectiveness of the eco-system approach highly depends on the initialization of new agents. In this paper we propose different initialization techniques, inspired from Deep Neural Network initialization and transfer learning, and study their impact.