Towards Probabilistic Question Answering Over Tabular Data

Abstract

Progress in question answering (QA) over tabular data has enabled reliable factual retrieval from relational tables. However, many real-world questions are probabilistic, requiring reasoning under uncertainty and latent conditional dependencies that are not directly stored in individual cells. We introduce LUCARIO, a large-scale benchmark for probabilistic QA over real-world tabular datasets, covering diverse domains, dependency patterns, and natural language variations. We propose Auto-BN, a fully automatic neuro-symbolic framework that induces a Bayesian Network from raw tables, translates natural-language questions into formal probabilistic queries, and performs exact inference. Experiments across multiple LLM backbones show that Auto-BN consistently outperforms NL2SQL, retrieval-based, and premise-based baselines, while maintaining a near-zero error rate. LUCARIO fills a relevant gap and provides a practical and scalable testbed for advancing probabilistic reasoning over structured data.

arXiv commentAccepted to AACL-IJCNLP 2026 (Main Conference). This version substantially revises and extends the original report and corresponds to the version accepted for publication. The author list has been updated to reflect the contributors to the present version