Using Large Language Models to Generate Authentic Multi-agent Knowledge Work Datasets

TL;DR

This paper presents a configurable system that uses Large Language Models to generate diverse, annotated, and context-rich multi-agent knowledge work datasets, addressing limitations of existing data collections.

Contribution

The authors introduce a novel multi-agent dataset generator leveraging LLMs to produce realistic, annotated knowledge work data with background context, enabling privacy-preserving evaluations.

Findings

53% of generated documents rated as realistic by human raters

74% of real documents rated as realistic by human raters

Analysis of authenticity criteria and improvement suggestions

Abstract

Current publicly available knowledge work data collections lack diversity, extensive annotations, and contextual information about the users and their documents. These issues hinder objective and comparable data-driven evaluations and optimizations of knowledge work assistance systems. Due to the considerable resources needed to collect such data in real-life settings and the necessity of data censorship, collecting such a dataset appears nearly impossible. For this reason, we propose a configurable, multi-agent knowledge work dataset generator. This system simulates collaborative knowledge work among agents producing Large Language Model-generated documents and accompanying data traces. Additionally, the generator captures all background information, given in its configuration or created during the simulation process, in a knowledge graph. Finally, the resulting dataset can be utilized and shared without privacy or confidentiality concerns. This paper introduces our approach's design and vision and focuses on generating authentic knowledge work documents using Large Language Models. Our study involving human raters who assessed 53% of the generated and 74% of the real documents as realistic demonstrates the potential of our approach. Furthermore, we analyze the authenticity criteria mentioned in the participants' comments and elaborate on potential improvements for identified common issues.