arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2008.00025cs.LGstat.ML
收藏

重新思考默认值:一种低成本、高效率的超参数定义策略

Rethinking Default Values: a Low Cost and Efficient Strategy to Define Hyperparameters

Federal Technology University - Paraná(巴拉那联邦技术大学);São Paulo State University (UNESP)(圣保罗州立大学);University of São Paulo (USP)(圣保罗大学);Federal University of Ouro Preto (UFOP)(欧鲁普雷图联邦大学);State University of Londrina (UEL)(隆德里纳州立大学)

Rafael Gomes Mantovani, André Luis Debiaso Rossi, Edesio Alcobaça, Jadson Castro Gertrudes, Sylvio Barbon Junior, André Carlos Ponce de Leon Ferreira de Carvalho

摘要:

机器学习(ML)算法已经越来越多地应用于几个不同领域的问题。尽管它们越来越受欢迎,但它们的预测性能通常受到为其超参数(HP)分配的值的影响。因此,研究人员和实践者面临着如何设定这些价值观的挑战。许多用户对ML算法及其HP值的影响了解有限,因此不会利用合适的设置。他们通常通过反复试验来定义HP值,这是非常主观的,不能保证找到好的值,并且依赖于用户体验。调整技术搜索HP值,能够最大限度地提高给定数据集的诱导模型的预测性能,但具有计算成本高的缺点。因此,从业者使用算法开发者或实现算法的工具建议的默认值。虽然默认值通常会导致模型具有可接受的预测性能,但同一算法的不同实现可能会建议不同的默认值。为了在调优和使用默认值之间保持平衡,我们提出了一种生成新的优化默认值的策略。我们的方法是基于一个小的优化值集,能够获得预测性能值比流行的工具提供的默认设置更好。在进行了一个大型实验和对结果的仔细分析后,我们得出结论,我们的方法提供了更好的默认值。此外,与调谐值相比,它导致了有竞争力的解决方案,使其更容易使用,成本更低。我们还提取了一些简单的规则来指导从业者决定是使用我们的新方法还是HP调优方法。

英文摘要:

Machine Learning (ML) algorithms have been increasingly applied to problems from several different areas. Despite their growing popularity, their predictive performance is usually affected by the values assigned to their hyperparameters (HPs). As consequence, researchers and practitioners face the challenge of how to set these values. Many users have limited knowledge about ML algorithms and the effect of their HP values and, therefore, do not take advantage of suitable settings. They usually define the HP values by trial and error, which is very subjective, not guaranteed to find good values and dependent on the user experience. Tuning techniques search for HP values able to maximize the predictive performance of induced models for a given dataset, but have the drawback of a high computational cost. Thus, practitioners use default values suggested by the algorithm developer or by tools implementing the algorithm. Although default values usually result in models with acceptable predictive performance, different implementations of the same algorithm can suggest distinct default values. To maintain a balance between tuning and using default values, we propose a strategy to generate new optimized default values. Our approach is grounded on a small set of optimized values able to obtain predictive performance values better than default settings provided by popular tools. After performing a large experiment and a careful analysis of the results, we concluded that our approach delivers better default values. Besides, it leads to competitive solutions when compared to tuned values, making it easier to use and having a lower cost. We also extracted simple rules to guide practitioners in deciding whether to use our new methodology or a HP tuning approach.

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。

↑