arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.07753cs.LG
收藏

不确定性门控生成建模

Uncertainty-Gated Generative Modeling

Electrical Engineering and Computer Science University of California, Berkeley(加州大学伯克利分校电子工程与计算机科学系);School of Electrical and Computer Engineering Cornell University(康奈尔大学电气与计算机工程学院)

Xingrui Gu, Haixi Zhang

AI总结:

本文提出UGGM方法,通过不确定性门控机制提升金融时间序列预测的鲁棒性和准确性,实验显示在NYISO数据集上MSE显著降低。

摘要:

金融时间序列预测是一个高风险问题,其中制度转变和冲击使得精确但过度自信的模型具有危险性。我们提出了不确定性门控生成建模(UGGM),该方法将不确定性视为内部控制信号,通过门控重参数化进行表示,通过相似性和置信度路由进行传播,通过不确定性控制的预测分布进行生成,同时结合由不确定性驱动的正则化和校准来抑制误校准。在弱创新自编码器(WIAE-GPF)上实例化,我们的UG-WIAE-GPF显著提高了风险敏感的预测能力,在NYISO上实现了63.5%的MSE降低(0.3508→0.1281),在冲击区间内表现出改进的鲁棒性(mSE:0.2739→0.1748)。

英文摘要:

Financial time-series forecasting is a high-stakes problem where regime shifts and shocks make point-accurate yet overconfident models dangerous. We propose Uncertainty-Gated Generative Modeling (UGGM), which treats uncertainty as an internal control signal that gates (i) representation via gated reparameterization, (ii) propagation via similarity and confidence routing, and (iii) generation via uncertainty-controlled predictive distributions, together with uncertainty-driven regularization and calibration to curb miscalibration. Instantiated on Weak Innovation AutoEncoder (WIAE-GPF), our UG-WIAE-GPF significantly improves risk-sensitive forecasting, delivering a 63.5\% MSE reduction on NYISO (0.3508 $\rightarrow$ 0.1281), with improved robustness under shock intervals (mSE: 0.2739 $\rightarrow$ 0.1748).

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。

↑