预测还是不预测？用于表格数据插补的比例掩码自编码器
To Predict or Not To Predict? Proportionally Masked Autoencoders for Tabular Data Imputation
AI总结：
本文针对表格数据插补中均匀随机掩码破坏缺失分布的问题，提出比例掩码自编码器（PMAE），通过保持缺失统计分布并采用高效MLP token混合，在多种缺失模式下取得更优插补性能。
摘要：
掩码自编码器（MAEs）最近在表格数据插补中展现了有效性。然而，由于表格数据固有的异质性，MAEs中常用的均匀随机掩码策略可能破坏缺失性分布，导致性能欠佳。为解决这一问题，我们提出了一种用于MAEs的比例掩码策略。具体来说，我们首先基于数据集中观测到的比例计算缺失性统计信息，然后生成与这些统计信息对齐的掩码，确保掩码后缺失性分布得以保留。此外，我们认为，与注意力机制相比，基于简单MLP的token混合在表格领域具有竞争力或通常更优的性能，同时计算效率更高，尤其是在存在固有异质性的表格领域中。实验结果验证了所提出的比例掩码策略在表格数据集各种缺失数据模式下的有效性。代码可在：\url{https://github.com/normal-kim/PMAE}获取。
英文摘要：
Masked autoencoders (MAEs) have recently demonstrated effectiveness in tabular data imputation. However, due to the inherent heterogeneity of tabular data, the uniform random masking strategy commonly used in MAEs can disrupt the distribution of missingness, leading to suboptimal performance. To address this, we propose a proportional masking strategy for MAEs. Specifically, we first compute the statistics of missingness based on the observed proportions in the dataset, and then generate masks that align with these statistics, ensuring that the distribution of missingness is preserved after masking. Furthermore, we argue that simple MLP-based token mixing offers competitive or often superior performance compared to attention mechanisms while being more computationally efficient, especially in the tabular domain with the inherent heterogeneity. Experimental results validate the effectiveness of the proposed proportional masking strategy across various missing data patterns in tabular datasets. Code is available at: \url{https://github.com/normal-kim/PMAE}.
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。
