pdfQA: Diverse, Challenging, and Realistic Question Answering over PDFs

TL;DR

This paper introduces pdfQA, a comprehensive dataset for question answering over PDFs, designed to evaluate diverse skills and challenges across multiple domains and complexity levels, aiding end-to-end system development.

Contribution

The paper presents pdfQA, a large, multi-domain, and complexity-diverse dataset for PDF question answering, including both real and synthetic data with detailed annotations.

Findings

LLMs struggle with complex PDF QA tasks

Dataset reveals challenges in information retrieval and parsing

Diverse complexity dimensions impact QA performance

Abstract

PDFs are the second-most used document type on the internet (after HTML). Yet, existing QA datasets commonly start from text sources or only address specific domains. In this paper, we present pdfQA, a multi-domain 2K human-annotated (real-pdfQA) and 2K synthetic dataset (syn-pdfQA) differentiating QA pairs in ten complexity dimensions (e.g., file type, source modality, source position, answer type). We apply and evaluate quality and difficulty filters on both datasets, obtaining valid and challenging QA pairs. We answer the questions with open-source LLMs, revealing existing challenges that correlate with our complexity dimensions. pdfQA presents a basis for end-to-end QA pipeline evaluation, testing diverse skill sets and local optimizations (e.g., in information retrieval or parsing).