SingularClip：防止谱崩溃以在持续学习和强化学习中保持可塑性
SingularClip: Preventing Spectral Collapse to Maintain Plasticity in Continual and Reinforcement Learning
浏览论文内容
中文总结 AI 辅助
SingularClip是一种定期裁剪权重矩阵奇异值的方法，可缓解训练中因奇异值各向异性增长导致的神经网络可塑性丧失问题，在两类学习任务上性能优于基线。
中文摘要 AI 辅助
在非平稳任务上训练的神经网络常丧失拟合新目标的能力，该现象被称为可塑性丧失。我们通过实证和理论分析发现，训练期间权重矩阵奇异值的各向异性不断增长是可塑性丧失的新来源。为缓解该问题，我们提出SingularClip，一种定期裁剪所有权重矩阵奇异值的方法。我们在持续监督学习和深度强化学习的一系列任务中，证明SingularClip的性能显著优于基线方法。
英文摘要
Neural networks trained on nonstationary tasks frequently lose the ability to fit new targets, a phenomenon referred to as loss of plasticity. We identify a novel source of plasticity loss due to the growing anisotropy of weight matrices' singular values during training, and analyze this phenomenon both empirically and theoretically. To mitigate this issue, we introduce SingularClip, a procedure that periodically clips the singular values of all weight matrices. We show that SingularClip performs strongly against baselines across a range of tasks in both continual supervised learning and deep reinforcement learning.
发表机构
- University of Toronto（多伦多大学）
- Vector Institute（向量研究所）
- Mila（米拉研究所）
- Polytechnique Montréal（蒙特利尔理工学院）
机构由 AI 辅助整理，请以论文原文为准。
