标题： 在时间分布变化下的最优经验风险最小化 显示英文标题

标题： Optimal Empirical Risk Minimization under Temporal Distribution Shifts

摘要： 时间分布的变化是机器学习模型在动态演变环境中训练和部署时面临的一个关键挑战。 本文介绍了RIDER（在动态演变制度下的风险最小化），它在时间分布变化下推导出最优加权的经验风险最小化过程。 我们的方法在随机分布变化模型中具有理论基础，其中随机变化是数据生成过程中的大量不可预测变化的叠加。 我们表明，常见的加权方案，如合并所有数据、指数加权数据以及仅使用最近的数据，自然地作为我们框架中的特例出现。 我们证明了当在Yearbook数据集上作为微调步骤应用时，RIDER在Wild-Time中的各种基准方法中始终提高了样本外预测性能。 此外，我们展示了RIDER在两个其他现实任务中优于标准加权策略：预测股票市场波动性和预测纽约市出租车数据中的乘车时间。 显示英文摘要

摘要： Temporal distribution shifts pose a key challenge for machine learning models trained and deployed in dynamically evolving environments. This paper introduces RIDER (RIsk minimization under Dynamically Evolving Regimes) which derives optimally-weighted empirical risk minimization procedures under temporal distribution shifts. Our approach is theoretically grounded in the random distribution shift model, where random shifts arise as a superposition of numerous unpredictable changes in the data-generating process. We show that common weighting schemes, such as pooling all data, exponentially weighting data, and using only the most recent data, emerge naturally as special cases in our framework. We demonstrate that RIDER consistently improves out-of-sample predictive performance when applied as a fine-tuning step on the Yearbook dataset, across a range of benchmark methods in Wild-Time. Moreover, we show that RIDER outperforms standard weighting strategies in two other real-world tasks: predicting stock market volatility and forecasting ride durations in NYC taxi data.