Corpus map improves search by linking related entities across documents
Follow the Entities: A Corpus Map for Agentic Search
Computation and LanguageArtificial IntelligenceInformation RetrievalMachine Learning
Summary
Sometimes when you look for answers in many documents, important clues are scattered in different places. The authors created a system called CorpusMap that organizes these documents by the people, places, or things they mention. This makes it easier for a computer to jump between related documents and find all the clues at once, instead of searching blindly. CorpusMap helps get better answers faster and uses fewer resources by remembering these links ahead of time.
What this means in practice
- •For enterprise search teams: Organize large company document collections by shared entities to enhance evidence gathering and improve answer quality in question answering systems.
- •For legal discovery teams: Connect related case documents through entity-based linking to find corroborating evidence efficiently across large legal datasets.
Authors
Soyeong Jeong, Sujay Kumar Jauhar, Sung Ju Hwang, Andrew Joohun Nam
Abstract
Answering questions and completing tasks over large document collections often requires connecting evidence spread across multiple documents, such as a project's approval recorded in one, its requirements in another, and its latest status in a third. Recent LLM agents approach this by iteratively searching the full corpus rather than reading only a fixed set of top-ranked documents. However, when the corpus is exposed only as a flat collection of files, a relevant document gives no indication of how it relates to others, so the agent must rediscover these relationships for every query, often missing complementary evidence while simultaneously consuming substantial additional tokens. To address this, we introduce CorpusMap, a navigation layer that organizes the corpus around its recurring entities, which are identifiable from the documents themselves and can link a single document to many others across sources. Specifically, CorpusMap represents each recurring entity as an Entity Page that aggregates information about it and links to every document that refers to it, forming a graph between entities and documents that the agent can traverse to gather otherwise disconnected evidence. Moreover, since CorpusMap is constructed offline by resolving mentions of the same entity across documents, its links are shared across queries rather than rediscovered repeatedly at inference time. Using 7 different models with 3 benchmark datasets, we show that CorpusMap improves both evidence discovery and answer quality over raw-corpus agentic search while using fewer tokens on average, and further outperforms 4 alternative navigation layers, suggesting that entities serve as effective anchors for navigating large document collections.