Knowledge Transfer in Deep Reinforcement Learning via an RL-Specific GAN-Based Correspondence Function
This addresses the challenge of knowledge reuse in deep reinforcement learning for AI systems, though it is incremental as it builds on existing GAN methods.
The paper tackled the problem of knowledge transfer and generalization in deep reinforcement learning by introducing a modified CycleGAN approach with new loss components, achieving 100% knowledge transfer in identical tasks and up to a 30% reduction in training time for rotated tasks in Atari Pong.
Deep reinforcement learning has demonstrated superhuman performance in complex decision-making tasks, but it struggles with generalization and knowledge reuse - key aspects of true intelligence. This article introduces a novel approach that modifies Cycle Generative Adversarial Networks specifically for reinforcement learning, enabling effective one-to-one knowledge transfer between two tasks. Our method enhances the loss function with two new components: model loss, which captures dynamic relationships between source and target tasks, and Q-loss, which identifies states significantly influencing the target decision policy. Tested on the 2-D Atari game Pong, our method achieved 100% knowledge transfer in identical tasks and either 100% knowledge transfer or a 30% reduction in training time for a rotated task, depending on the network architecture. In contrast, using standard Generative Adversarial Networks or Cycle Generative Adversarial Networks led to worse performance than training from scratch in the majority of cases. The results demonstrate that the proposed method ensured enhanced knowledge generalization in deep reinforcement learning.