论文标题
优化的可重复性：理论框架和限制
Reproducibility in Optimization: Theoretical Framework and Limits
论文作者
论文摘要
我们启动了优化可重复性的正式研究。我们在面对嘈杂或容易出错的操作（例如不精确或随机梯度计算或不精确初始化）的情况下定义了优化程序可重复性的定量度量。然后，我们分析了感兴趣的几个凸优化设置，例如平滑，非平滑和强烈的目标函数，并在每种环境中的可重复性范围内建立紧密的界限。我们的分析揭示了计算和可重复性之间的基本权衡：需要更多计算（足够），以便更好地可重复性。
We initiate a formal study of reproducibility in optimization. We define a quantitative measure of reproducibility of optimization procedures in the face of noisy or error-prone operations such as inexact or stochastic gradient computations or inexact initialization. We then analyze several convex optimization settings of interest such as smooth, non-smooth, and strongly-convex objective functions and establish tight bounds on the limits of reproducibility in each setting. Our analysis reveals a fundamental trade-off between computation and reproducibility: more computation is necessary (and sufficient) for better reproducibility.
