用联合分布瓦瑟斯坦距离提升时序预测的分布对齐效果
DistDF: Time-Series Forecasting Needs Joint-Distribution Wasserstein Alignment
- 提出基于联合分布瓦瑟斯坦距离的新对齐方法
- 在多个数据集上显著优于传统均方误差方法
- 适合需要精准分布建模的时序预测任务
训练时序预测模型需对齐模型预测的条件分布与真实标签序列的分布。标准直接预测（DF）方法通常通过最小化条件负对数似然来实现，常用均方误差估计。然而当标签序列存在自相关时，该估计会产生偏差。本文提出DistDF，通过最小化预测序列与标签序列条件分布间的分布差异来实现对齐。由于有限时序观测下条件差异难以估计，我们引入一种时序预测的联合分布瓦瑟斯坦差异，可严格上界目标条件差异。该差异度量可计算、可微分，且兼容梯度优化。大量实验表明，DistDF能提升多种预测模型性能，并达到领先水平。代码已公开于 https://anonymous.4open.science/r/DistDF-F66B。
原文摘要 · Abstract (English)
Training time-series forecasting models requires aligning the conditional distribution of model forecasts with that of the label sequence. The standard direct forecast (DF) approach resorts to minimizing the conditional negative log-likelihood, typically estimated by the mean squared error. However, this estimation proves biased when the label sequence exhibits autocorrelation. In this paper, we propose DistDF, which achieves alignment by minimizing a distributional discrepancy between the conditional distributions of forecast and label sequences. Since such conditional discrepancies are difficult to estimate from finite time-series observations, we introduce a joint-distribution Wasserstein discrepancy for time-series forecasting, which provably upper bounds the conditional discrepancy of interest. The proposed discrepancy is tractable, differentiable, and readily compatible with gradient-based optimization. Extensive experiments show that DistDF improves diverse forecasting models and achieves leading performance. Code is available at https://anonymous.4open.science/r/DistDF-F66B.
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。