Abstract

Despite the transformative impact of deep learning on text, audio, and imagedatasets, its dominance in tabular data, especially in the medical domain wheredata are often scarce, remains less clear. In this paper, we propose X2Graph, anovel deep learning method that achieves strong performance on small biologicaltabular datasets. X2Graph leverages external knowledge about the relationshipsbetween table columns, such as gene interactions, to convert each sample into agraph structure. This transformation enables the application of standardmessage passing algorithms for graph modeling. Our X2Graph method demonstratessuperior performance compared to existing tree-based and deep learning methodsacross three cancer subtyping datasets.

 

Quick Read (beta)