Localized TabICLv2: Scaling Tabular In-Context Learning through k-NN

arXiv:2608.16429v1 Announce Type: new Abstract: Foundational models for tabular data have made significant progress in recent years, with TabICLv2 reporting state-of-the-art performance on several tabular classification tasks. However, full-context tabular ICL still suffers from attention cost that grows with the training-context size, which limits its ability to handle large datasets efficiently. Localized TabICLv2 introduces a method that reduces the inference cost of TabICLv2 by retrieving only the k nearest training neighbours for each test point, measured by similarity in the model's Stage 2 row-representation space, rather than using the full training context. This requires no architectural changes, and we show that accuracy retention can be improved through additional Stage 2 and Stage 3 fine-tuning. On TabArena classification tasks, the fine-tuned localized model retains 98.64% of Full TabICLv2 accuracy and it achieves a median 2.18imes speedup in batch inference, and reaches approximately 249imes median speedup in the single-query serving setting.

Source: arXiv cs.LG | 2026-08-18