动态TMoE：一种针对非平稳时间序列预测的漂移感知动态专家混合框架
Dynamic TMoE: A Drift-Aware Dynamic Mixture of Experts Framework for Non-Stationary Time Series Forecasting
AI总结：
本文提出Dynamic TMoE框架，通过动态构建异构专家和剪枝冗余专家来优化容量，并利用时间记忆路由器确保稳定且上下文感知的专家选择，从而在非平稳时间序列预测中实现更优性能。
摘要：
非平稳时间序列预测面临由演变分布偏移带来的挑战，静态模型难以捕捉这些变化。虽然混合专家（MoE）架构提供了解耦复杂漂移模式的有前景范式，但现有方法受限于固定专家池和无记忆路由，阻碍了其适应突发制度转变的能力。为此，我们提出Dynamic TMoE框架，将架构进化与时间连续性统一在学习阶段。通过最大均值偏差（MMD）检测分布偏移，动态实例化异构专家并剪枝冗余专家以优化容量。此外，时间记忆路由器利用循环状态和异常库确保稳定、上下文感知的专家选择，无需测试时更新。在九个基准测试中的实验表明，该方法实现了最先进的性能，将MSE减少10.4%，MAE减少7.8%。代码可在https://github.com/andone-07/Dynamic-TMoE获取。
英文摘要：
Non-stationary time series forecasting is challenged by evolving distribution shifts that static models struggle to capture. While Mixture-of-Experts (MoE) architectures offer a promising paradigm for decoupling complex drift patterns, existing approaches are limited by fixed expert pools and memoryless routing, hampering their ability to adapt to abrupt regime shifts. To address this, we propose Dynamic TMoE, a framework that unifies architectural evolution with temporal continuity during learning phase. By detecting distribution shifts via Maximum Mean Discrepancy (MMD), we dynamically instantiate heterogeneous experts and prune redundant ones to optimize capacity. Additionally, a temporal memory router leverages recurrent states and an anomaly repository to ensure stable, context-aware expert selection without requiring test-time updates. Experiments on nine benchmarks demonstrate state-of-the-art performance, reducing MSE by 10.4% and MAE by 7.8%. Code is available at https://github.com/andone-07/Dynamic-TMoE.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
