Data Skeptic Process Mining with LLMs
7 snips
Sep 24, 2024 
David Obembe, a recent graduate from the University of Tartu, dives into his master's thesis on blending large language models with process mining tools. He explains how process mining uses event logs to map out inefficiencies in business processes. Fascinating insights include the evolution of these techniques post-LLM integration, enhancing data retrieval and insights. David shares his experiments with Retrieval Augmented Generation and discusses the challenges of prompt engineering, highlighting the balance between accuracy and model reliability.
AI Snips
Chapters
Transcript
Episode notes
Process Mining Tools
- Process mining tools convert event logs into process maps.
- These maps help analysts find bottlenecks and inefficiencies.
LLMs in Process Mining
- LLMs enhance process mining by directly answering questions about processes.
- Two approaches exist: direct prompting with table data and SQL queries.
Prompt Engineering
- Prompt engineering is crucial for accurate LLM results in process mining.
- David Obembe iteratively refined prompts, adding definitions like "bottleneck."