LLM-Based Process Name Generation for Subprocess Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional process mining tools lack the ability to assign intuitive and descriptive names to subprocesses and variants, making it difficult to analyze and refer to specific portions of processes.
Innovation Solution
Utilizing large language models to generate names for subprocesses and variants based on textual descriptions of process models, allowing for iterative naming and annotation of process model portions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional process mining tools assign numbers to variants, then the process mining functionality is maintained, but the intuitiveness and understandability of variant identification deteriorates
Solution Approach 1:
The patent changes the parameter of variant identification from numerical identifiers to natural language descriptions generated by LLMs. This transformation maintains the unique identification capability while adding semantic meaning, allowing users to understand the functional characteristics of each variant directly from its name.
Solution Approach 2:
The patent introduces LLMs as an intermediary between the process mining tool and the user. The LLM generates descriptive names based on the sequence of activities in each variant, acting as a mediator that translates complex process data into human-understandable identifiers without losing functional information.
2Ease of operation
If conventional process mining tools do not provide names to subprocesses, then the system complexity is reduced, but the ease of referencing and analyzing subprocesses deteriorates
Solution Approach 1:
The patent applies preliminary action by generating names for subprocesses and variants before the user needs to reference them. The LLM pre-processes the process model data to create meaningful identifiers, so when users need to analyze or reference subprocesses, the naming task has already been completed, improving ease of operation without adding complexity to the user interaction.
3Ease of operation
If LLMs are used to generate process names, then the intuitiveness and descriptive quality of names is improved, but the processing time and computational resources increase
Solution Approach 1:
The patent applies partial action by generating names for only the most important or frequently referenced subprocesses and variants, rather than attempting to name every single element. This selective approach maintains high descriptive quality for critical process elements while reducing the overall processing time and computational burden.
Data Source
AI summary
Systems and methods for generating names for portions, such as, e.g., subprocesses or variants, of a process model are provided. One or more prompts defining 1) instructions, 2) a textual description of a process model of a process, and 3) one or more portions of the process model are received. A name for each of the one or more portions of the process model is generated using a large language model based on the instructions. The name for each of the one or more portions of the process model is output.