Retrieval & Fine-Tuning for In-Context Tabular Models
This addresses the problem of handling diverse and complex tabular data for machine learning practitioners, representing an incremental improvement over existing in-context learning methods.
The paper tackled the challenge of scaling transformer-based in-context learning to larger and more complex tabular datasets by proposing a retrieval and fine-tuning method, resulting in LoCalPFN which achieved state-of-the-art performance on 95 datasets, significantly boosting performance over the base model.
Tabular data is a pervasive modality spanning a wide range of domains, and the inherent diversity poses a considerable challenge for deep learning. Recent advancements using transformer-based in-context learning have shown promise on smaller and less complex datasets, but have struggled to scale to larger and more complex ones. To address this limitation, we propose a combination of retrieval and fine-tuning: we can adapt the transformer to a local subset of the data by collecting nearest neighbours, and then perform task-specific fine-tuning with this retrieved set of neighbours in context. Using TabPFN as the base model -- currently the best tabular in-context learner -- and applying our retrieval and fine-tuning scheme on top results in what we call a locally-calibrated PFN, or LoCalPFN. We conduct extensive evaluation on 95 datasets curated by TabZilla from OpenML, upon which we establish a new state-of-the-art with LoCalPFN -- even with respect to tuned tree-based models. Notably, we show a significant boost in performance compared to the base in-context model, demonstrating the efficacy of our approach and advancing the frontier of deep learning in tabular data.