Tag
NVIDIA 发布预训练关系型基础模型 Kumo Relational，可直接在企业数据库上通过单次前向传播完成预测，无需训练与特征工程；在 41 项任务与 15 个数据库上超越监督式关系深度学习及专家特征的梯度提升树管线。
The paper proposes GenScript, an inference-only synthetic data pipeline that replaces per-dataset model training by passing a deterministic statistical profile of the source data to an LLM, which infers field semantics and integrity constraints that a coding agent then compiles into an auditable sampler. The approach unifies tabular, temporal, and relational data synthesis, building generators in minutes while closely matching leading methods in fidelity and preserving key integrity constraints.
This paper proposes RelShap, a framework that incorporates relational constraints and data provenance into Shapley value computation, making explanations more faithful to the data-generating process. It is estimator-agnostic and composes with existing SHAP estimators while exploiting functional dependencies to reduce runtime.
This paper introduces AutoGrable, a method that scores candidate graph constructions for tabular data without training a GNN, using a label-alignment risk based on 1-WL color refinement, enabling cheap search for effective table-to-graph mappings.
GRAB uses a GNN encoder to convert relational tables into latent tokens for frozen LLMs, achieving significant performance gains in multi-table question answering.
This paper introduces RelAD, a reconstruction-based framework for detecting anomalies in relational databases by jointly modeling attribute and relational edge reconstruction. Extensive experiments on six new benchmarks show RelAD outperforms existing methods.