Causality Learning With Wasserstein Generative Adversarial Networks-猫眼课题宝

DOI：

10.5121/ijaia.2022.13301

发表时间：

2022-06

期刊：

ArXiv

影响因子：

通讯作者：

中科院分区：

其他

文献类型：

作者：

H. Petkov;Colin Hanley;Feng Dong

由于组合搜索空间的存在，从数据中学习因果结构的传统方法面临着巨大的挑战。最近，该问题被描述为一个带无圈约束的连续优化框架来学习有向无环图(DAG)。这种框架允许使用深度生成模型进行因果结构学习，以更好地捕获数据样本分布和DAG之间的关系。然而，到目前为止，还没有研究将沃瑟斯坦距离用于因果结构学习。我们的模型DAG-WGAN结合了基于瓦瑟斯坦的对抗性损失和自动编码器体系结构中的无循环约束。它在提高数据生成能力的同时学习因果结构。我们将DAG-WGAN的性能与其他不涉及Wasserstein度量的模型进行了比较，以确定其对因果结构学习的贡献。实验表明，当基数较高时，我们的模型具有更好的性能。

Conventional methods for causal structure learning from data face significant challenges due to combinatorial search space. Recently, the problem has been formulated into a continuous optimization framework with an acyclicity constraint to learn Directed Acyclic Graphs (DAGs). Such a framework allows the utilization of deep generative models for causal structure learning to better capture the relations between data sample distributions and DAGs. However, so far no study has experimented with the use of Wasserstein distance in the context of causal structure learning. Our model named DAG-WGAN combines the Wasserstein-based adversarial loss with an acyclicity constraint in an auto-encoder architecture. It simultaneously learns causal structures while improving its data generation capability. We compare the performance of DAG-WGAN with other models that do not involve the Wasserstein metric in order to identify its contribution to causal structure learning. Our model performs better with high cardinality data according to our experiments.

Causality Learning With Wasserstein Generative Adversarial Networks

文献摘要

相似文献