通过筛选有效时间步，提升深度时序模型预测精度。
Selective Learning for Deep Time Series Forecasting
- 仅对可信时间步计算损失，避开噪声与异常点
- 在8个真实数据集上，最高降低37.4%的均方误差
- 适合高噪声场景下的时序预测任务
得益于捕捉复杂时序模式的强大能力，深度学习显著推进了时序预测（TSF）的发展。然而，由于时序数据对噪声和异常值敏感，深度模型易出现严重过拟合。当前主流方法对所有时间步统一使用均方误差（MSE）损失进行优化，无法区分不确定或异常的时间步，导致过拟合。为此，我们提出一种面向深度时序预测的新型选择性学习策略：在优化过程中仅选取部分时间步计算MSE损失，引导模型聚焦于可泛化的趋势，忽略不可靠部分。该框架引入双重掩码机制：(1) 不确定性掩码，利用残差熵筛选不确定性高的时间步；(2) 异常掩码，通过残差下界估计剔除异常时间步。在8个真实世界数据集上的大量实验表明，该方法能显著提升典型先进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%。
原文摘要 · Abstract (English)
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.
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。