Causal Factor Inference for Process Abnormality Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for deriving auxiliary information to guide appropriate actions from machine learning results lack mechanisms to provide insights into the causal relationships between variables, making it difficult for non-experts to take accurate and timely actions.
Innovation Solution
A factor inference device and method that utilize a knowledge model representing causal relationships between events in a process, combined with operational data to infer and present the factors contributing to a phenomenon, including abnormality indices and causal routes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning models are used for abnormality detection and prediction, then prediction accuracy and abnormality detection capability are improved, but the ability to provide actionable insights and causal understanding deteriorates
Solution Approach 1:
The patent introduces an explanation model as an intermediary between the machine learning model and the user. This explanation model generates natural language descriptions that bridge the gap between accurate predictions and human-understandable causal insights, allowing users to understand why predictions were made without compromising prediction accuracy
Solution Approach 2:
The system segments the information processing into distinct components: the machine learning model handles prediction accuracy while the explanation model handles causal interpretation. This segmentation allows each component to specialize in its strength without compromising the other
2Measurement precision
If deep knowledge about analysis target is required for accurate inference, then inference accuracy is improved, but ease of operation deteriorates
Solution Approach 1:
The explanation model acts as a mediator that translates complex technical concepts and causal relationships into natural language that non-experts can understand. This allows users without deep domain knowledge to access accurate inference results and understand the reasoning behind them
Solution Approach 2:
The system provides self-service by automatically generating explanations and insights without requiring users to possess specialized knowledge. The explanation model autonomously interprets the machine learning model's reasoning and presents it in an accessible format
3Measurement precision
If comprehensive analysis of multiple monitoring items is performed, then abnormality detection accuracy is improved, but processing time and computational load increase
Solution Approach 1:
The explanation model extracts only the most relevant causal relationships and insights from the comprehensive analysis results, presenting them in a condensed and easily digestible format. This extraction process maintains the benefits of comprehensive analysis while reducing the time and cognitive load required to interpret results
Data Source
AI summary
A factor inference device for inferring factor of a phenomenon in a process includes: a knowledge model acquisition unit configured to acquire a knowledge model having an event occurring in the process as a node and expressed in a network form connecting a plurality of the nodes to each other, with respect to a causal relationship of the phenomenon in the process; an information creation unit configured to create information including at least an abnormality index related to the event based on data collected from the process; a data combining unit configured to associate a node of the knowledge model with the information corresponding to the node; and a factor inference unit configured to infer and present the factor of the phenomenon based on a structure of the knowledge model and the information associated with the node.