arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2402.09891cs.LGstat.ML
收藏

因果预测器能否更好地泛化到新领域?

Do causal predictors generalize better to new domains?

Max Planck Institute for Intelligent Systems(马克斯·普朗克智能系统研究所);Tübingen AI Center(蒂宾根人工智能中心);Max Planck ETH Center for Learning Systems(马克斯·普朗克ETH学习系统中心)

Vivian Y. Nastl, Moritz Hardt

AI总结:

本文通过16个跨领域表格预测任务实证检验因果特征是否有助于领域泛化,发现因果预测器在域内外准确率及跨域性能下降方面均不优于使用全部特征的模型,并验证了现有因果域泛化方法和因果发现算法同样未能带来提升。

摘要:

我们研究在因果特征上训练的机器学习模型跨领域泛化的能力。我们考虑了表格数据集上的16个预测任务,涵盖健康、就业、教育、社会福利和政治等应用领域。每个数据集都包含多个领域,使我们能够测试在一个领域上训练的模型在另一个领域上的表现。对于每个预测任务,我们选择对预测目标具有因果影响的特征。我们的目标是检验这样一个假设:在因果特征上训练的模型能够更好地跨领域泛化。无一例外,我们发现使用所有可用特征的预测器,无论其是否具有因果性,其域内和域外准确率都优于使用因果特征的预测器。此外,即使是从一个领域到另一个领域的绝对准确率下降幅度,因果预测器也不比使用所有特征的模型更好。另外,我们表明,最近的用于领域泛化的因果机器学习方法在我们的评估中并不比在因果特征集上训练的标准预测器表现更好。同样,因果发现算法要么无法运行,要么选出的因果变量表现并不优于我们的选择。广泛的稳健性检验确认,我们的发现在变量错误分类下依然稳定。

英文摘要:

We study how well machine learning models trained on causal features generalize across domains. We consider 16 prediction tasks on tabular datasets covering applications in health, employment, education, social benefits, and politics. Each dataset comes with multiple domains, allowing us to test how well a model trained in one domain performs in another. For each prediction task, we select features that have a causal influence on the target of prediction. Our goal is to test the hypothesis that models trained on causal features generalize better across domains. Without exception, we find that predictors using all available features, regardless of causality, have better in-domain and out-of-domain accuracy than predictors using causal features. Moreover, even the absolute drop in accuracy from one domain to the other is no better for causal predictors than for models that use all features. In addition, we show that recent causal machine learning methods for domain generalization do not perform better in our evaluation than standard predictors trained on the set of causal features. Likewise, causal discovery algorithms either fail to run or select causal variables that perform no better than our selection. Extensive robustness checks confirm that our findings are stable under variable misclassification.

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。

↑