New Long-Horizon Q-Learning method improves reinforcement learning accuracy

Researchers have introduced Long-Horizon Q-Learning (LQL), a novel method designed to improve the stability of value-based reinforcement learning. LQL addresses the issue of compounding estimation errors in traditional Q-learning by incorporating an inequality-based backstop. This approach penalizes violations of bounds derived from observed action sequences, using existing network outputs to avoid additional computational overhead. Experiments show LQL consistently outperforms standard 1-step and n-step TD learning across various benchmarks. AI

影响 Introduces a method to stabilize reinforcement learning, potentially improving agent performance in complex, long-horizon tasks.

排序理由 This is a research paper detailing a new algorithm for reinforcement learning. [lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

• Long-Horizon Q-Learning
• n-step TD learning
• Q-learning
• reinforcement learning
• TD updates

• 论文
• 其他

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →