用粒子滤波解决深度学习中的顺序依赖问题，防止遗忘和塑性丧失。
Permutation Invariant Learning with High-Dimensional Particle Filters
- 基于高维粒子滤波实现训练顺序无关的参数更新
- 在多个持续学习任务上性能优于传统方法，方差更低
- 适合需要稳定长期学习的场景，如持续监督与强化学习
深度模型的序列学习常因梯度算法的顺序依赖性导致灾难性遗忘和可塑性下降。本文提出一种基于高维粒子滤波的新型排列不变学习框架。理论证明粒子滤波对训练小批量或任务的顺序具有不变性，可有效缓解遗忘与塑性丧失。我们设计了高效优化高维模型的粒子滤波方法，融合贝叶斯推理与梯度优化优势。在连续监督与强化学习基准（包括SplitMNIST、SplitCIFAR100、ProcGen）上的大量实验表明，该方法性能更优且方差显著降低。
原文摘要 · Abstract (English)
Sequential learning in deep models often suffers from challenges such as catastrophic forgetting and loss of plasticity, largely due to the permutation dependence of gradient-based algorithms, where the order of training data impacts the learning outcome. In this work, we introduce a novel permutation-invariant learning framework based on high-dimensional particle filters. We theoretically demonstrate that particle filters are invariant to the sequential ordering of training minibatches or tasks, offering a principled solution to mitigate catastrophic forgetting and loss-of-plasticity. We develop an efficient particle filter for optimizing high-dimensional models, combining the strengths of Bayesian methods with gradient-based optimization. Through extensive experiments on continual supervised and reinforcement learning benchmarks, including SplitMNIST, SplitCIFAR100, and ProcGen, we empirically show that our method consistently improves performance, while reducing variance compared to standard baselines.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。