Papers by Cameron Barrie
Satyrn: A Platform for Analytics Augmented Generation (2024.emnlp-main)
Copied to clipboard
Marko Sterbentz, Cameron Barrie, Shubham Shahi, Abhratanu Dutta, Donna Hooshmand, Harper Pack, Kristian Hammond
| Challenge: | Large language models (LLMs) can generate fluent and coherent language, but not all information can be retrieved from text. |
| Approach: | They propose an approach that leverages structured data to generate fact sets that are converted to text and passed to an LLM. |
| Outcome: | The proposed approach generates reports in which over 86% of claims are accurate while maintaining high levels of fluency and coherence. |