面向深度时间序列预测的选择性学习
Selective Learning for Deep Time Series Forecasting
AI总结：
针对深度时间序列预测中因统一优化所有时间步导致的过拟合问题，提出含双掩码机制的选择性学习策略，筛选泛化性时间步优化，在8个数据集上显著提升多个主流模型预测性能。
摘要：
得益于捕捉复杂时间模式的强大能力，深度学习（DL）已极大推动了时间序列预测（TSF）的发展。然而，由于时间序列本身易受噪声和异常值影响，深度模型往往存在严重的过拟合问题。当前主流的深度学习范式通过均方误差（MSE）损失对所有时间步进行统一优化，无差别地学习那些不确定和异常的时间步，最终导致过拟合。为解决这一问题，我们提出了一种面向深度时间序列预测的新型选择性学习策略。具体而言，选择性学习会筛选全部时间步的一个子集来计算优化过程中的MSE损失，引导模型聚焦于具有泛化性的时间步，同时忽略非泛化性的时间步。我们的框架引入双掩码机制来定位目标时间步：（1）不确定性掩码，利用残差熵过滤不确定的时间步；（2）异常掩码，采用残差下界估计排除异常时间步。在8个真实世界数据集上开展的大量实验表明，选择性学习可显著提升典型前沿深度模型的预测性能，其中Informer的MSE降低37.4%，TimesNet降低8.4%，iTransformer降低6.5%。
英文摘要：
Benefiting from high capacity for capturing complex temporal patterns, deep learning (DL) has significantly advanced time series forecasting (TSF). However, deep models tend to suffer from severe overfitting due to the inherent vulnerability of time series to noise and anomalies. The prevailing DL paradigm uniformly optimizes all timesteps through the MSE loss and learns those uncertain and anomalous timesteps without difference, ultimately resulting in overfitting. To address this, we propose a novel selective learning strategy for deep TSF. Specifically, selective learning screens a subset of the whole timesteps to calculate the MSE loss in optimization, guiding the model to focus on generalizable timesteps while disregarding non-generalizable ones. Our framework introduces a dual-mask mechanism to target timesteps: (1) an uncertainty mask leveraging residual entropy to filter uncertain timesteps, and (2) an anomaly mask employing residual lower bound estimation to exclude anomalous timesteps. Extensive experiments across eight real-world datasets demonstrate that selective learning can significantly improve the predictive performance for typical state-of-the-art deep models, including 37.4% MSE reduction for Informer, 8.4% for TimesNet, and 6.5% for iTransformer.
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。
