arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.12564cs.LGcs.NAmath.NAphysics.flu-dynstat.ML
收藏

通过使用低差异序列进行训练来提高深度学习算法的准确性

Enhancing accuracy of deep learning algorithms by training with low-discrepancy sequences

ETH Zürich(苏黎世联邦理工学院)

Siddhartha Mishra, T. Konstantin Rusch

摘要:

我们提出了一种基于低差异序列作为训练集的深度监督学习算法。通过理论论证和大量的数值实验相结合,我们证明,该算法显着优于标准的深度学习算法,是基于随机选择的训练数据,在中等高的维度的问题。该算法为科学计算中的许多底层地图提供了一种有效的替代方法。

英文摘要:

We propose a deep supervised learning algorithm based on low-discrepancy sequences as the training set. By a combination of theoretical arguments and extensive numerical experiments we demonstrate that the proposed algorithm significantly outperforms standard deep learning algorithms that are based on randomly chosen training data, for problems in moderately high dimensions. The proposed algorithm provides an efficient method for building inexpensive surrogates for many underlying maps in the context of scientific computing.

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。

↑