arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.08166cs.LGcs.AIstat.ML
收藏

超参数优化的自动终止

Automatic Termination for Hyperparameter Optimization

ETH Zürich(苏黎世联邦理工学院);Amazon Web Services(亚马逊云科技)

Anastasia Makarova, Huibin Shen, Valerio Perrone, Aaron Klein, Jean Baptiste Faddoul, Andreas Krause, Matthias Seeger, Cedric Archambeau

摘要:

贝叶斯优化(BO)是机器学习中超参数优化(HPO)的一种常用方法。在其核心,BO迭代地评估有希望的配置,直到用户定义的预算(如挂钟时间或迭代次数)耗尽。虽然调优后的最终性能在很大程度上取决于提供的预算,但很难预先指定最佳值。在这项工作中,我们提出了一个有效的和直观的终止标准BO自动停止的过程,如果它是足够接近的全局最优。我们的关键见解是,真实目标(测试数据的预测性能)和可计算目标(验证性能)之间的差异表明,一旦优化目标的次优性被统计估计误差所主导,就应该停止。在广泛的范围内的现实世界的HPO问题和基线,我们表明,我们的终止标准实现了更好的权衡之间的测试性能和优化时间。此外,我们发现过拟合可能发生在HPO的背景下,这可以说是一个被忽视的问题,在文献中,并显示我们的终止标准如何有助于减轻这种现象在小型和大型数据集。

英文摘要:

Bayesian optimization (BO) is a widely popular approach for the hyperparameter optimization (HPO) in machine learning. At its core, BO iteratively evaluates promising configurations until a user-defined budget, such as wall-clock time or number of iterations, is exhausted. While the final performance after tuning heavily depends on the provided budget, it is hard to pre-specify an optimal value in advance. In this work, we propose an effective and intuitive termination criterion for BO that automatically stops the procedure if it is sufficiently close to the global optimum. Our key insight is that the discrepancy between the true objective (predictive performance on test data) and the computable target (validation performance) suggests stopping once the suboptimality in optimizing the target is dominated by the statistical estimation error. Across an extensive range of real-world HPO problems and baselines, we show that our termination criterion achieves a better trade-off between the test performance and optimization time. Additionally, we find that overfitting may occur in the context of HPO, which is arguably an overlooked problem in the literature, and show how our termination criterion helps to mitigate this phenomenon on both small and large datasets.

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。

↑