2 papers
stat.ML2026
PACE: Plug-and-Play Contextual Embedding for Feature Screening with Pretrained Tabular Foundation Models
Qi Qin, Erbo Li, Ting Wei +4
In high-dimensional tabular learning, feature screening provides a lightweight, model-agnostic way to remove irrelevant features before model fitting. However, scoring raw values d…
cs.LG2026
GEAR: Generative Expansion and Real Anchoring for Two-Stage Distillation of Tabular Foundation Models
Qi Qin, Jiajie Zhu, Dali Chen +6
Tabular foundation models (TFMs) achieve strong performance through in-context learning, but context-dependent inference imposes substantial latency and memory costs, hindering lar…