Towards Probabilistic Question Answering Over Tabular Data

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Shen, Chen, Rahman, Sajjadur, Hruschka, Estevam
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915359380996096
author Shen, Chen
Rahman, Sajjadur
Hruschka, Estevam
author_facet Shen, Chen
Rahman, Sajjadur
Hruschka, Estevam
contents Current approaches for question answering (QA) over tabular data, such as NL2SQL systems, perform well for factual questions where answers are directly retrieved from tables. However, they fall short on probabilistic questions requiring reasoning under uncertainty. In this paper, we introduce a new benchmark LUCARIO and a framework for probabilistic QA over large tabular data. Our method induces Bayesian Networks from tables, translates natural language queries into probabilistic queries, and uses large language models (LLMs) to generate final answers. Empirical results demonstrate significant improvements over baselines, highlighting the benefits of hybrid symbolic-neural reasoning.
format Preprint
id arxiv_https___arxiv_org_abs_2506_20747
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards Probabilistic Question Answering Over Tabular Data
Shen, Chen
Rahman, Sajjadur
Hruschka, Estevam
Computation and Language
68T50, 68T37
I.2.7
Current approaches for question answering (QA) over tabular data, such as NL2SQL systems, perform well for factual questions where answers are directly retrieved from tables. However, they fall short on probabilistic questions requiring reasoning under uncertainty. In this paper, we introduce a new benchmark LUCARIO and a framework for probabilistic QA over large tabular data. Our method induces Bayesian Networks from tables, translates natural language queries into probabilistic queries, and uses large language models (LLMs) to generate final answers. Empirical results demonstrate significant improvements over baselines, highlighting the benefits of hybrid symbolic-neural reasoning.
title Towards Probabilistic Question Answering Over Tabular Data
topic Computation and Language
68T50, 68T37
I.2.7
url https://arxiv.org/abs/2506.20747