Fine-Tuning the Retrieval Mechanism for Tabular Deep Learning
This work addresses the problem of improving deep learning performance on tabular data for researchers and practitioners, though it appears incremental as it builds on existing pretraining and retrieval concepts.
The paper tackles the performance gap between deep learning and tree-based models in tabular data by exploring a retrieval mechanism, finding that fine-tuning the pretrained TabPFN model with this approach notably surpasses existing methods.
While interests in tabular deep learning has significantly grown, conventional tree-based models still outperform deep learning methods. To narrow this performance gap, we explore the innovative retrieval mechanism, a methodology that allows neural networks to refer to other data points while making predictions. Our experiments reveal that retrieval-based training, especially when fine-tuning the pretrained TabPFN model, notably surpasses existing methods. Moreover, the extensive pretraining plays a crucial role to enhance the performance of the model. These insights imply that blending the retrieval mechanism with pretraining and transfer learning schemes offers considerable potential for advancing the field of tabular deep learning.