Newest papers

On Data Scaling in Tabular ML: More Data, Better Generalization?

Ivan Rubachev, Nikita Lazarev, Artem Babenko · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26794 · cited by 0 works (OpenAlex) · open-access PDF

Arabic

وقد أظهرت هذه العملية الأخيرة في دراسة البيانات المتطورة أهمية إعدادات التقييم المختلفة والواقعية. ومع ذلك، لا تزال البيانات الكبيرة (1M+) تحت الاستكشاف. لتطوير تركيز البيانات المتطورة، فإننا نجمع ثمانية مجموعات البيانات المتطورة في العالم الحقيقي مع ما يقرب من عشرة ملايين مثال للتقييم، مع الحفاظ على تغيرات التوزيع الطبيعي للمتطلبات واستخدام التقييمات المتطورة بشكل أكثر تعقيداً حتى أثناء تغيرات التوزيع. ومع ذلك، فإن نموذجات GBDT وMLP تظهر تركيزات التقييم المزدوجة على نطاق واسع، مع مزايا من كل من البيانات التدريبية الإضافية والتحسينات الفوريّة. قد تظهر

Machine translation (Meta M2M100); the English original follows.

Abstract

Recent work in tabular machine learning has emphasized the importance of diverse and realistic evaluation settings. The niche of large-scale (1M+) data, however, remains under-explored. To study training-data scaling, we assemble eight real-world tabular datasets with up to around ten million training examples, retaining the natural distribution shifts of the tasks and using random-split evaluation as a control. We find that GBDTs and MLP-based models show broadly similar scaling trends, with gains from both additional training data and algorithmic improvements. Tabular foundation models show more varied scaling behavior, with recent versions benefiting more consistently from larger contexts even under distribution shift. However, scaling training data under distribution shift introduces new challenges: in-distribution and out-of-distribution performance mostly improve together at smaller scales but can diverge at larger scales. We examine this through MLP capacity sweeps and comparisons of the validation distributions used for tuning and early stopping in MLPs and GBDTs. We also compare common practices of using validation data, contrasting its use for model selection in MLP pipelines with using it for in-context fitting in tabular foundation model pipelines. Under stronger shifts, MLPs selected using validation representative of the evaluation conditions generally maintain or improve performance as training data grows. In contrast, including such examples in context improves foundation-model performance, but does not prevent deterioration with increasing training size in our experiments.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Scaling, Computer science, Context (archaeology), Machine learning, Selection (genetic algorithm), Model selection

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange