A Deep Hierarchical Approach to Lifelong Learning in Minecraft
This addresses the problem of efficient knowledge retention and transfer in high-dimensional lifelong learning for AI agents in complex environments like Minecraft, representing an incremental improvement with novel techniques.
The paper tackles lifelong learning in Minecraft by proposing a hierarchical deep reinforcement learning network that reuses and transfers knowledge through deep skill networks and skill distillation, achieving superior performance and lower sample complexity compared to Deep Q Networks in sub-domains.
We propose a lifelong learning system that has the ability to reuse and transfer knowledge from one task to another while efficiently retaining the previously learned knowledge-base. Knowledge is transferred by learning reusable skills to solve tasks in Minecraft, a popular video game which is an unsolved and high-dimensional lifelong learning problem. These reusable skills, which we refer to as Deep Skill Networks, are then incorporated into our novel Hierarchical Deep Reinforcement Learning Network (H-DRLN) architecture using two techniques: (1) a deep skill array and (2) skill distillation, our novel variation of policy distillation (Rusu et. al. 2015) for learning skills. Skill distillation enables the HDRLN to efficiently retain knowledge and therefore scale in lifelong learning, by accumulating knowledge and encapsulating multiple reusable skills into a single distilled network. The H-DRLN exhibits superior performance and lower learning sample complexity compared to the regular Deep Q Network (Mnih et. al. 2015) in sub-domains of Minecraft.