Agent-centric Information Access

TL;DR

This paper proposes a scalable framework for agent-centric information access, enabling dynamic selection and querying of specialized LLMs as knowledge agents to improve domain-specific information retrieval.

Contribution

It introduces a novel framework for dynamically ranking and querying expert LLMs, addressing challenges in expert inference, selection, and response synthesis in a scalable manner.

Findings

Developed a scalable evaluation framework using retrieval-augmented generation.

Implemented clustering techniques to manage thousands of specialized models.

Demonstrated potential to scale to millions of expert models.

Abstract

As large language models (LLMs) become more specialized, we envision a future where millions of expert LLMs exist, each trained on proprietary data and excelling in specific domains. In such a system, answering a query requires selecting a small subset of relevant models, querying them efficiently, and synthesizing their responses. This paper introduces a framework for agent-centric information access, where LLMs function as knowledge agents that are dynamically ranked and queried based on their demonstrated expertise. Unlike traditional document retrieval, this approach requires inferring expertise on the fly, rather than relying on static metadata or predefined model descriptions. This shift introduces several challenges, including efficient expert selection, cost-effective querying, response aggregation across multiple models, and robustness against adversarial manipulation. To address these issues, we propose a scalable evaluation framework that leverages retrieval-augmented generation and clustering techniques to construct and assess thousands of specialized models, with the potential to scale toward millions.