KoGuN: Accelerating Deep Reinforcement Learning via Integrating Human Suboptimal Knowledge

TL;DR

KoGuN is a framework that integrates human suboptimal prior knowledge into reinforcement learning, significantly improving learning efficiency especially with limited or imperfect prior knowledge.

Contribution

This paper introduces KoGuN, a novel end-to-end framework combining fuzzy rule-based human knowledge with RL, enhancing learning speed and efficiency.

Findings

Achieves faster learning in discrete and continuous control tasks.

Effective even with low-quality human prior knowledge.

Outperforms baseline RL algorithms in sample efficiency.

Abstract

Reinforcement learning agents usually learn from scratch, which requires a large number of interactions with the environment. This is quite different from the learning process of human. When faced with a new task, human naturally have the common sense and use the prior knowledge to derive an initial policy and guide the learning process afterwards. Although the prior knowledge may be not fully applicable to the new task, the learning process is significantly sped up since the initial policy ensures a quick-start of learning and intermediate guidance allows to avoid unnecessary exploration. Taking this inspiration, we propose knowledge guided policy network (KoGuN), a novel framework that combines human prior suboptimal knowledge with reinforcement learning. Our framework consists of a fuzzy rule controller to represent human knowledge and a refine module to fine-tune suboptimal prior knowledge. The proposed framework is end-to-end and can be combined with existing policy-based reinforcement learning algorithm. We conduct experiments on both discrete and continuous control tasks. The empirical results show that our approach, which combines human suboptimal knowledge and RL, achieves significant improvement on learning efficiency of flat RL algorithms, even with very low-performance human prior knowledge.