Enregistré dans:
Détails bibliographiques
Auteurs principaux: Huang, Xingyue, Tichelman, Louis, Kim, Jinwoo, Olejniczak, Krzysztof, Ceylan, İsmail İlkan
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2605.07840
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910201687310336
author Huang, Xingyue
Tichelman, Louis
Kim, Jinwoo
Olejniczak, Krzysztof
Ceylan, İsmail İlkan
author_facet Huang, Xingyue
Tichelman, Louis
Kim, Jinwoo
Olejniczak, Krzysztof
Ceylan, İsmail İlkan
contents Relational learning is a challenging problem that has motivated a wide range of approaches, including graph-based models (e.g., graph neural networks, graph transformers), tabular methods (e.g., tabular foundation models), and sequence-based approaches (e.g., large language models), each with its own advantages and limitations. We propose RelAgent, an LLM-based autonomous data scientist for relational learning, which operates in two phases. In the search phase, an LLM agent uses database, validation, and evaluation workspace tools to construct SQL feature programs and select a predictive model. In the inference phase, the resulting program is executed without further LLM calls. The final predictor consists of SQL queries and a classical model, enabling fast, deterministic, and intrinsically interpretable predictions: features are human-readable queries, and predictions depend only on the resulting query-defined feature map, enabling scalable deployment using standard database systems.
format Preprint
id arxiv_https___arxiv_org_abs_2605_07840
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle RelAgent: LLM Agents as Data Scientists for Relational Learning
Huang, Xingyue
Tichelman, Louis
Kim, Jinwoo
Olejniczak, Krzysztof
Ceylan, İsmail İlkan
Machine Learning
Relational learning is a challenging problem that has motivated a wide range of approaches, including graph-based models (e.g., graph neural networks, graph transformers), tabular methods (e.g., tabular foundation models), and sequence-based approaches (e.g., large language models), each with its own advantages and limitations. We propose RelAgent, an LLM-based autonomous data scientist for relational learning, which operates in two phases. In the search phase, an LLM agent uses database, validation, and evaluation workspace tools to construct SQL feature programs and select a predictive model. In the inference phase, the resulting program is executed without further LLM calls. The final predictor consists of SQL queries and a classical model, enabling fast, deterministic, and intrinsically interpretable predictions: features are human-readable queries, and predictions depend only on the resulting query-defined feature map, enabling scalable deployment using standard database systems.
title RelAgent: LLM Agents as Data Scientists for Relational Learning
topic Machine Learning
url https://arxiv.org/abs/2605.07840