LLM Activity Mapping for Meaning-Aware Process Mining

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current process mining technologies rely on manual activity mapping, which is time-consuming and requires significant knowledge and experience, and similarity measures that focus on activity labels rather than their actual meanings.

Innovation Solution

The use of large language models to determine mappings between activities extracted from a process model and those executed during instances of process execution, by receiving prompts with instructions and activities, and outputting mappings such as one-to-one, one-to-many, or many-to-one.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual activity mapping is used, then mapping accuracy can be maintained through expert knowledge, but time consumption and operational complexity increase significantly

Engineering Contradiction:
Improvemapping accuracyVSAvoidtime consumption
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent introduces an intermediary system comprising a process model translator and activity mapper that automatically translates process model definitions into executable mapping rules. This intermediary layer bridges the gap between manual expert knowledge and automated processing, enabling accurate activity mapping without requiring direct human intervention for each mapping decision.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary actions by pre-translating process model definitions into structured formats and pre-establishing mapping rules before actual activity mapping is needed. This advance preparation stores mapping logic in reusable formats, eliminating the need for time-consuming manual mapping during execution while maintaining expert-level accuracy.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If manual activity mapping is used, then expert knowledge can ensure quality mappings, but the operational difficulty and expertise requirements increase

Engineering Contradiction:
Improvemapping qualityVSAvoidoperational difficulty
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The system implements self-service by automatically generating activity mappings through the activity mapper component, which autonomously translates process model definitions into mapping rules without requiring expert operators. The system serves itself by maintaining and updating mapping logic through automated translation processes, eliminating the need for human expertise in operational mapping tasks.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical system of manual expert analysis with an automated computational system. The process model translator and activity mapper use algorithmic processes to analyze process definitions and generate mappings, substituting human cognitive mechanics with automated information processing that maintains reliability while dramatically improving ease of operation.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If similarity measures based on activity labels are used, then automated mapping can be achieved, but the ability to capture actual activity meaning is lost

Engineering Contradiction:
Improveautomation capabilityVSAvoidactivity meaning accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The system changes the parameters used for mapping by transitioning from simple activity label comparison to comprehensive process model definition analysis. Instead of relying solely on surface-level label similarity, the system analyzes detailed process definitions, translations, and contextual information to capture the actual meaning of activities, thereby maintaining measurement precision while preserving automation capability.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent adds another dimension to the mapping process by incorporating process model translations and detailed definitions beyond simple activity labels. This dimensional expansion allows the system to analyze the semantic meaning and contextual context of activities, enabling automated mapping that captures actual activity meaning rather than relying solely on label similarity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentEP4530769A1Activity mapping for process mining using large language models
Publication Date: 2025.04.02 UIPATH INC
  • EP4530769A1 patent drawingFigure 1
  • EP4530769A1 patent drawingFigure 2A
  • EP4530769A1 patent drawingFigure 2B

AI summary

Systems and methods for determining a mapping between activities are provided. One or more prompts defining 1) instructions and 2) activities executed during one or more instances of execution of a process are received. A mapping between one or more activities extracted from a process model of the process and one or more of the activities executed during the one or more instances of execution is determined using a large language model based on the instructions. The mapping is output.