好、坏与丑：以专家知识增强黑盒模型
机器学习
2020-05-12 v2 机器学习
摘要
我们在材料科学的背景下解决非唯一参数拟合问题。具体而言，我们提出通过以两个不同层次的专家知识增强黑盒人工神经网络（ANN）模型来解决参数空间中的歧义，并将其与纯黑盒模型进行基准比较。
引用
@article{arxiv.1907.11105,
title = {The Good, the Bad and the Ugly: Augmenting a black-box model with expert knowledge},
author = {Raoul Heese and Michał Walczak and Lukas Morand and Dirk Helm and Michael Bortz},
journal= {arXiv preprint arXiv:1907.11105},
year = {2020}
}
备注
International Conference on Artificial Neural Networks (ICANN) 2019