Abstract

This paper explores the application of Large Language Models (LLMs) for page stream segmentation in insurance document automation. We investigate parameter-efficient fine-tuning approaches and analyze calibration challenges that arise in high-stakes applications where incorrect segmentation can have significant business impact.

Key Contributions

  • Novel approach: Applied LLMs to page stream segmentation for insurance documents
  • Parameter-efficient fine-tuning: Demonstrated effective adaptation methods for domain-specific tasks
  • Calibration analysis: Examined model confidence and reliability in high-stakes applications
  • Industry application: Real-world deployment considerations for insurance document processing

Impact

This work bridges the gap between academic LLM research and practical industry applications, particularly in the insurance sector where document processing automation can significantly improve efficiency while maintaining accuracy requirements.

Citation

@inproceedings{heidenreich2025page,
  title={Page Stream Segmentation with LLMs: Challenges and Applications in Insurance Document Automation},
  author={Heidenreich, Hunter and Dalvi, Ratish and Verma, Nikhil and Getachew, Yosheb},
  booktitle={Proceedings of the 31st International Conference on Computational Linguistics: Industry Track},
  pages={305--317},
  year={2025}
}