SkillTFM: Gated Skill Evolution for Training-Free Adaptation of Tabular Foundation Models

2026-08-06

AI Summary

This work addresses the challenge that tabular foundation models struggle to adapt efficiently under distribution shifts, semantic heterogeneity, and task-specific patterns without costly fine-tuning. To overcome this limitation, we propose SkillTFM, a training-free adaptation framework that uniquely integrates boundary evidence identification with gated skill evolution. By leveraging a verifiable and extensible skill library, SkillTFM enables cross-task knowledge reuse, transforming model adaptation from parameter updating into dynamic retrieval and composition of predefined skills. The framework is compatible with diverse tabular foundation models and demonstrates substantial performance gains on both simulated boundary scenarios and real-world electricity price forecasting tasks, improving the nonlinear boundary AUC from 0.699 to 0.898 and achieving overall AUC gains of 0.128–0.142, thereby validating its effectiveness and generalizability.

Abstract

Tabular data are ubiquitous in real-world applications and are crucial for data-driven prediction and decision-making across science, industry, finance, healthcare, and public services. Tabular foundation models (TFMs) have emerged as a promising paradigm for general-purpose tabular learning, offering reusable predictors across diverse datasets and substantially reducing the need for task-specific training, tuning, and model development. However, their practical deployment remains constrained by distribution shifts, heterogeneous feature semantics, and task-specific patterns that are difficult to capture without costly fine-tuning or additional labeled data. To this end, we propose SkillTFM, a training-free system that shifts TFM adaptation from parameter updates to the gated evolution of agentic skills. The core of SkillTFM is a verifiable and extensible skill bank that couples boundary evidence identification with gated skill evolution: the former characterizes task structure and base-model failure patterns, whereas the latter retrieves and extends reusable skills subject to explicit validation. Across simulated boundary settings and real-world electricity-price forecasting, SkillTFM improves AUC by 0.128--0.142, raises nonlinear-boundary AUC from 0.699 to 0.898. Furthermore, experiments across TFM backbones demonstrate the effectiveness and generality of SkillTFM.

Problem

• Tabular Foundation Models
• distribution shifts
• heterogeneous feature semantics
• training-free adaptation
• task-specific patterns

Innovation

• training-free adaptation
• tabular foundation models
• gated skill evolution
• skill bank
• distribution shift