Robust Hybrid Learning With Expert Augmentation
Robust Hybrid Learning With Expert Augmentation
复制标题
DOI：
--
复制
发表时间：
2022-02
期刊：
影响因子：
--
通讯作者：
Antoine Wehenkel;Jens Behrmann;Hsiang Hsu;G. Sapiro;Gilles Louppe and;J. Jacobsen
中科院分区：
文献类型：
--
作者：
Antoine Wehenkel;Jens Behrmann;Hsiang Hsu;G. Sapiro;Gilles Louppe and;J. Jacobsen
混合建模通过将专家模型与从数据中学习的机器学习（ML）组件相结合，减少了专家模型的错误指定。与许多ML算法类似，混合模型的性能保证仅限于训练分布。利用专家模型通常是有效的，即使在训练域之外的洞察力，我们克服了这一限制，引入了一种混合数据增强策略，称为\textit{专家增强}。基于混合建模的概率形式化，我们证明了专家增强，它可以被纳入现有的混合动力系统，提高泛化。我们经验验证专家增强三个控制实验建模动力系统与普通和偏微分方程。最后，我们评估潜在的现实世界的适用性专家增强数据集上的一个真实的双摆。
Hybrid modelling reduces the misspecification of expert models by combining them with machine learning (ML) components learned from data. Similarly to many ML algorithms, hybrid model performance guarantees are limited to the training distribution. Leveraging the insight that the expert model is usually valid even outside the training domain, we overcome this limitation by introducing a hybrid data augmentation strategy termed \textit{expert augmentation}. Based on a probabilistic formalization of hybrid modelling, we demonstrate that expert augmentation, which can be incorporated into existing hybrid systems, improves generalization. We empirically validate the expert augmentation on three controlled experiments modelling dynamical systems with ordinary and partial differential equations. Finally, we assess the potential real-world applicability of expert augmentation on a dataset of a real double pendulum.