Scalability of CausalMixFT versus GAN-based Augmentation in Tabular Foundation Model Fine-tuning

Fine-tuning tabular foundation models (TFMs) under data scarcity is challenging, as early stopping on even scarcer validation data often fails to capture true generalization performance. We propose CausalMixFT, a method that enhances fine-tuning robustness and downstream performance by generating structurally consistent synthetic samples using Structural Causal Models (SCMs) fitted on the target dataset. This approach augments limited real data with causally informed synthetic examples, preserving feature dependencies while expanding training diversity. Evaluated across 33 classification datas

Research goal: How does the scalability of CausalMixFT compare to GAN-based augmentation in fine-tuning tabular foundation models across varying dataset sizes in the TabMWM benchmark, evaluated in terms of validation accuracy and generalization gap?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.7/10.