Reasoning Factual Knowledge in Structured Data with Large Language Models

TL;DR

This paper introduces StructFact, a benchmark to evaluate large language models' ability to reason with structured data, revealing current limitations and guiding future improvements in factual knowledge inference.

Contribution

The paper presents StructFact, a comprehensive benchmark for assessing LLMs' reasoning over structured data, highlighting their current shortcomings and fostering advancements in knowledge reasoning.

Findings

LLMs show limited ability to infer factual knowledge from structured data.

The benchmark covers diverse tasks, domains, and regions for comprehensive evaluation.

Current training strategies do not fully address the challenges of reasoning with structured data.

Abstract

Large language models (LLMs) have made remarkable progress in various natural language processing tasks as a benefit of their capability to comprehend and reason with factual knowledge. However, a significant amount of factual knowledge is stored in structured data, which possesses unique characteristics that differ from the unstructured texts used for pretraining. This difference can introduce imperceptible inference parameter deviations, posing challenges for LLMs in effectively utilizing and reasoning with structured data to accurately infer factual knowledge. To this end, we propose a benchmark named StructFact, to evaluate the structural reasoning capabilities of LLMs in inferring factual knowledge. StructFact comprises 8,340 factual questions encompassing various tasks, domains, timelines, and regions. This benchmark allows us to investigate the capability of LLMs across five factual tasks derived from the unique characteristics of structural facts. Extensive experiments on a set of LLMs with different training strategies reveal the limitations of current LLMs in inferring factual knowledge from structured data. We present this benchmark as a compass to navigate the strengths and weaknesses of LLMs in reasoning with structured data for knowledge-sensitive tasks, and to encourage advancements in related real-world applications. Please find our code at https://github.com/EganGu/StructFact.