Aggregation Queries over Unstructured Text: Benchmark and Agentic Method

2026-02-01

AI Summary

This work addresses the challenge of ensuring completeness—i.e., retrieving all relevant results rather than just one—in aggregating queries over unstructured text. Focusing on large-scale corpora, the paper formally defines the task of entity-level aggregation queries and introduces AGGBench, the first benchmark specifically designed to evaluate completeness in such settings. To tackle this problem, the authors propose DFA, a modular agent framework that decomposes query processing into three interpretable stages: disambiguation, filtering, and aggregation. Experimental results demonstrate that DFA significantly outperforms strong retrieval-augmented generation (RAG) baselines and existing agent-based approaches in terms of evidence coverage, thereby substantially improving the completeness of aggregation query results.

Technology Category

• Data Mining & Knowledge Management: Intelligent Query Processing
• Natural Language Processing: Question Answering
• Multiagent Systems: Distributed Problem Solving

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Aggregation query over free text is a long-standing yet underexplored problem. Unlike ordinary question answering, aggregate queries require exhaustive evidence collection and systems are required to"find all,"not merely"find one."Existing paradigms such as Text-to-SQL and Retrieval-Augmented Generation fail to achieve this completeness. In this work, we formalize entity-level aggregation querying over text in a corpus-bounded setting with strict completeness requirement. To enable principled evaluation, we introduce AGGBench, a benchmark designed to evaluate completeness-oriented aggregation under realistic large-scale corpus. To accompany the benchmark, we propose DFA (Disambiguation--Filtering--Aggregation), a modular agentic baseline that decomposes aggregation querying into interpretable stages and exposes key failure modes related to ambiguity, filtering, and aggregation. Empirical results show that DFA consistently improves aggregation evidence coverage over strong RAG and agentic baselines. The data and code are available in \href{https://anonymous.4open.science/r/DFA-A4C1}.

Problem

• aggregation query
• unstructured text
• completeness
• entity-level aggregation
• evidence collection

Innovation

• aggregation query
• completeness
• AGGBench
• modular agentic method
• evidence coverage