Generative Adversarial Networks for Conformance Score Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional predictive data analysis solutions face inefficiencies and reliability issues in generating accurate conformance scores, particularly in complex business processes, due to the requirement of a prior model for conformance checking and the inability to account for the interplay between event trace and case-level attributes.
Innovation Solution
The use of generative adversarial neural networks with attention layers to perform conformance checking without a prior model, enabling the incorporation of complex rules and providing a conformance score based on state-level and attribute-level attention weights, allowing for predictive inferences from distinctions between core and secondary event features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional predictive data analysis solutions use prior models for conformance checking, then conformance determination can be performed, but the system complexity increases and reliability decreases due to the inability to account for interplay between event trace and case-level attributes
Solution Approach 1:
The patent segments the conformance checking task into two independent processing streams: state processing (using recurrent neural networks to capture temporal patterns in event traces) and attribute processing (using machine learning models to handle case-level attributes). This segmentation allows each component to specialize in its specific data type, improving overall reliability while managing complexity through modular architecture.
Solution Approach 2:
The patent introduces attention weight vectors as intermediaries that bridge the state processing and attribute processing streams. These attention mechanisms allow the system to dynamically weigh the importance of different event states and attributes when generating conformance scores, enabling the system to account for interplay between event trace and case-level attributes without requiring a single complex unified model.
2Measurement precision
If the system processes all event encoding data objects to generate comprehensive conformance scores, then predictive accuracy improves, but processing time increases
Solution Approach 1:
The patent applies partial action through the attention mechanism, which selectively processes only the most relevant portions of the event encoding data objects. The attention weight vectors identify and emphasize critical event states and attributes while downweighting less relevant information, allowing the system to achieve high measurement precision by focusing computational resources on the most impactful data elements rather than uniformly processing all data.
Solution Approach 2:
The patent implements local quality by applying different processing depths and attention mechanisms to different parts of the data. The state processing stream uses recurrent neural networks to capture temporal dependencies in specific event states, while the attribute processing stream handles case-level attributes with appropriate machine learning models. This localized processing approach ensures that each data element receives the appropriate level of processing attention, improving precision while managing processing time.
Data Source
AI summary
Various embodiments of the present invention provide methods, apparatus, systems, computing devices, computing entities, and/or the like for performing risk score generation predictive data analysis. Certain embodiments of the present invention utilize systems, methods, and computer program products that perform risk conformance mining predictive data analysis by utilizing machine learning frameworks that include state processing machine learning models and attribute processing machine learning models, where the machine learning frameworks may be trained as part of generative adversarial machine learning frameworks.