Goal-Aware Cross-Entropy for Multi-Target Reinforcement Learning
This addresses the challenge of sample inefficiency and poor generalization in multi-target environments for reinforcement learning agents, though it appears incremental as it builds on existing self-supervised and attention-based techniques.
The paper tackles the problem of multi-target reinforcement learning by proposing goal-aware cross-entropy loss and goal-discriminative attention networks, showing that GDAN outperforms state-of-the-art methods in task success ratio, sample efficiency, and generalization on visual navigation and robot arm manipulation tasks.
Learning in a multi-target environment without prior knowledge about the targets requires a large amount of samples and makes generalization difficult. To solve this problem, it is important to be able to discriminate targets through semantic understanding. In this paper, we propose goal-aware cross-entropy (GACE) loss, that can be utilized in a self-supervised way using auto-labeled goal states alongside reinforcement learning. Based on the loss, we then devise goal-discriminative attention networks (GDAN) which utilize the goal-relevant information to focus on the given instruction. We evaluate the proposed methods on visual navigation and robot arm manipulation tasks with multi-target environments and show that GDAN outperforms the state-of-the-art methods in terms of task success ratio, sample efficiency, and generalization. Additionally, qualitative analyses demonstrate that our proposed method can help the agent become aware of and focus on the given instruction clearly, promoting goal-directed behavior.