Hybrid Intelligence Event Detection Using Personalized Knowledge Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current AI/ML systems lack human knowledge and common sense, limiting their ability to understand event influences and provide adaptive interactions with users, necessitating a method to encapsulate human understanding for improved user experiences.
Innovation Solution
A method involving graph analysis to obtain subpopulation data associated with user personality and demographic characteristics, inferring psychological traits, and performing outcome linkage analysis to generate personalized knowledge graphs for event detection, monitoring, and anomaly detection in event data streams.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If AI/ML systems use traditional data analysis methods, then processing speed is improved, but understanding of human knowledge and common sense deteriorates
Solution Approach 1:
The patent merges symbolic AI (knowledge graphs representing human knowledge and common sense) with machine learning models (neural networks for pattern recognition). This combination allows the system to process data quickly through ML while simultaneously incorporating human knowledge through knowledge graphs, resolving the contradiction between processing speed and understanding of human knowledge.
Solution Approach 2:
The patent introduces knowledge graphs as an intermediary layer between raw data and ML model predictions. These knowledge graphs encode human knowledge, common sense, and domain expertise, serving as a mediator that guides and contextualizes ML processing, thereby preserving human understanding while maintaining computational efficiency.
2Quantity of substance
If AI/ML systems process large amounts of data, then analytical capability is improved, but adaptability to individual users deteriorates
Solution Approach 1:
The patent applies local quality by customizing knowledge graphs for individual users based on their specific characteristics, preferences, and contexts. Instead of using a single generic knowledge graph, the system creates user-specific instances that locally adapt the general knowledge base to each user's needs, thereby maintaining adaptability while processing large volumes of data.
Solution Approach 2:
The patent implements dynamic knowledge graphs that can be updated, modified, and personalized in real-time based on user interactions and new information. This dynamic nature allows the system to adapt to individual users while continuously processing large amounts of data, resolving the contradiction between data volume processing and user-specific adaptability.
3Device complexity
If AI/ML systems use generic models, then development complexity is reduced, but measurement precision of user-specific events deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-populating knowledge graphs with domain knowledge, common sense, and contextual information before deploying ML models. This pre-processing of knowledge structures reduces the complexity of model development while improving event detection accuracy, as the ML models can focus on pattern recognition rather than learning basic domain knowledge from scratch.
Data Source
AI summary
A method for event detection includes: obtaining a subpopulation data from a graph structure for performing a graph analysis, wherein the subpopulation data is associated with personality and demographic characteristics of users; obtaining a user profile associated with a target user; inferring psychological traits of the user by performing the graph analysis based on the user profile and the subpopulation data; performing an outcome linkage analysis based on labeled event outcome profiles and the inferred psychological traits to generate personalized knowledge graph data associated with the target user; and profiling, monitoring, or performing an anomaly detection for event data streams based on the personalized knowledge graph data.