研究人员开发了Dynamic TMoE，一个旨在改进非平稳时间序列预测的新型框架。该方法通过动态调整专家池并整合用于路由的时间记忆，解决了现有专家混合（MoE）模型的局限性。该系统使用最大均值差异（MMD）来实例化和修剪专家，从而优化模型容量，从而检测分布变化。实验表明，Dynamic TMoE在九个基准测试中取得了最先进的成果，显著降低了均方误差（MSE）和平均绝对误差（MAE）。 AI
影响 增强了时间序列预测能力，可能改进金融、天气和需求预测等应用。
排序理由 发布了关于新机器学习框架的学术论文。
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
arXiv:2605.25166v1 Announce Type: cross Abstract: Time series forecasting models are increasingly scaled through large Transformer backbones, yet most existing approaches process all series through a shared dense computation path despite substantial heterogeneity in temporal stru…
arXiv:2509.15105v3 Announce Type: replace Abstract: Time series forecasting (TSF) is critical in domains like energy, finance, healthcare, and logistics, requiring models that generalize across diverse datasets. Large pre-trained models such as Chronos and Time-MoE show strong ze…
arXiv:2605.20678v1 Announce Type: cross Abstract: 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 patt…
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 exp…