Not all datasets are born equal: On heterogeneous tabular data and adversarial examples

作者：

Highlights：

• Attacks on tabular data ignore complex features (nominal) and feature correlations.

• Mathematically define a valid real-world heterogeneous adversarial example.

• Use embedding function to preserve feature correlations and value consistency.

• Implement and evaluate the framework in three data domains and learning models.

摘要

•Attacks on tabular data ignore complex features (nominal) and feature correlations.•Mathematically define a valid real-world heterogeneous adversarial example.•Use embedding function to preserve feature correlations and value consistency.•Implement and evaluate the framework in three data domains and learning models.

论文关键词：Adversarial examples,Adversarial learning,Tabular data

论文评审过程：Received 31 August 2021, Revised 24 January 2022, Accepted 3 February 2022, Available online 14 February 2022, Version of Record 19 February 2022.

论文官网地址：https://doi.org/10.1016/j.knosys.2022.108377